{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8aa8ee84-b4d7-49e7-8cc1-65a6c4da7bb3",
   "metadata": {},
   "source": [
    "# Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "47fc8147-2146-4dd7-8e95-154cda123c0f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:30.585861Z",
     "iopub.status.busy": "2025-01-29T21:54:30.585373Z",
     "iopub.status.idle": "2025-01-29T21:54:33.848403Z",
     "shell.execute_reply": "2025-01-29T21:54:33.847809Z",
     "shell.execute_reply.started": "2025-01-29T21:54:30.585837Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[codecarbon WARNING @ 13:54:33] Multiple instances of codecarbon are allowed to run at the same time.\n",
      "[codecarbon INFO @ 13:54:33] [setup] RAM Tracking...\n",
      "[codecarbon INFO @ 13:54:33] [setup] CPU Tracking...\n",
      "[codecarbon WARNING @ 13:54:33] We saw that you have a Apple M3 Pro but we don't know it. Please contact us.\n",
      "[codecarbon WARNING @ 13:54:33] No CPU tracking mode found. Falling back on CPU constant mode. \n",
      " Mac OS and ARM processor detected: Please enable PowerMetrics sudo to measure CPU\n",
      "\n",
      "[codecarbon WARNING @ 13:54:33] We saw that you have a Apple M3 Pro but we don't know it. Please contact us.\n",
      "[codecarbon INFO @ 13:54:33] CPU Model on constant consumption mode: Apple M3 Pro\n",
      "[codecarbon WARNING @ 13:54:33] No CPU tracking mode found. Falling back on CPU constant mode.\n",
      "[codecarbon INFO @ 13:54:33] [setup] GPU Tracking...\n",
      "[codecarbon INFO @ 13:54:33] No GPU found.\n",
      "[codecarbon INFO @ 13:54:33] >>> Tracker's metadata:\n",
      "[codecarbon INFO @ 13:54:33]   Platform system: macOS-15.2-arm64-arm-64bit-Mach-O\n",
      "[codecarbon INFO @ 13:54:33]   Python version: 3.13.1\n",
      "[codecarbon INFO @ 13:54:33]   CodeCarbon version: 3.0.0_rc0\n",
      "[codecarbon INFO @ 13:54:33]   Available RAM : 18.000 GB\n",
      "[codecarbon INFO @ 13:54:33]   CPU count: 12\n",
      "[codecarbon INFO @ 13:54:33]   CPU model: Apple M3 Pro\n",
      "[codecarbon INFO @ 13:54:33]   GPU count: None\n",
      "[codecarbon INFO @ 13:54:33]   GPU model: None\n",
      "[codecarbon INFO @ 13:54:33] Saving emissions data to file /Users/andrebach/code/frugal-ai-submission/emissions.csv\n"
     ]
    }
   ],
   "source": [
    "import asyncio\n",
    "import textwrap\n",
    "from datetime import datetime\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from datasets import load_dataset\n",
    "from sklearn.metrics import accuracy_score\n",
    "from transformers import AutoTokenizer\n",
    "\n",
    "from tasks.text import DESCRIPTIONS, ROUTE, bert_model\n",
    "from tasks.utils.emissions import clean_emissions_data, get_space_info, tracker\n",
    "from tasks.utils.evaluation import TextEvaluationRequest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "076a669f-8bde-425e-b9ae-321d6023ba60",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:33.849394Z",
     "iopub.status.busy": "2025-01-29T21:54:33.849194Z",
     "iopub.status.idle": "2025-01-29T21:54:33.854360Z",
     "shell.execute_reply": "2025-01-29T21:54:33.853840Z",
     "shell.execute_reply.started": "2025-01-29T21:54:33.849370Z"
    }
   },
   "outputs": [],
   "source": [
    "%config InlineBackend.figure_format='retina'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81715352-7cc6-41cf-8df4-7180c950b042",
   "metadata": {},
   "source": [
    "# Modified code from text.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f382c570-af16-404e-9d0b-818e7c8d52d6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:33.855661Z",
     "iopub.status.busy": "2025-01-29T21:54:33.855453Z",
     "iopub.status.idle": "2025-01-29T21:54:33.858301Z",
     "shell.execute_reply": "2025-01-29T21:54:33.857940Z",
     "shell.execute_reply.started": "2025-01-29T21:54:33.855642Z"
    }
   },
   "outputs": [],
   "source": [
    "MODEL_TYPE = \"bert-mini\"\n",
    "# Define the label mapping\n",
    "LABEL_MAPPING = {\n",
    "    \"0_not_relevant\": 0,\n",
    "    \"1_not_happening\": 1,\n",
    "    \"2_not_human\": 2,\n",
    "    \"3_not_bad\": 3,\n",
    "    \"4_solutions_harmful_unnecessary\": 4,\n",
    "    \"5_science_unreliable\": 5,\n",
    "    \"6_proponents_biased\": 6,\n",
    "    \"7_fossil_fuels_needed\": 7,\n",
    "}\n",
    "LABEL_MAPPING_INV = {v: k for k, v in LABEL_MAPPING.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1f777cab-cc30-4654-a441-0f8045f50ee8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:33.859121Z",
     "iopub.status.busy": "2025-01-29T21:54:33.858932Z",
     "iopub.status.idle": "2025-01-29T21:54:33.863608Z",
     "shell.execute_reply": "2025-01-29T21:54:33.862752Z",
     "shell.execute_reply.started": "2025-01-29T21:54:33.859105Z"
    }
   },
   "outputs": [],
   "source": [
    "def fetch_test_dataset(request, data_version):\n",
    "    # This commit\n",
    "    # https://huggingface.co/datasets/QuotaClimat/frugalaichallenge-text-train/commit/4748e7b548154d188a377a96e104b9b31fcd45b4\n",
    "    # seemingly changed the way the data is organized and labeled as train & test.\n",
    "    # data_version = \"old\" maintains the original code for splitting\n",
    "    # \"new\" takes the newly so-called test set\n",
    "    # Which seems to be the same length, but a different randomization, as the old. :(\n",
    "\n",
    "    # Load and prepare the dataset\n",
    "    dataset = load_dataset(request.dataset_name)\n",
    "    # Convert string labels to integers\n",
    "    dataset = dataset.map(lambda x: {\"label\": LABEL_MAPPING[x[\"label\"]]})\n",
    "\n",
    "    if data_version == \"old\":\n",
    "        # Split dataset\n",
    "        train_test = dataset[\"train\"].train_test_split(\n",
    "            test_size=request.test_size, seed=request.test_seed\n",
    "        )\n",
    "        test_dataset = train_test[\"test\"]\n",
    "        return test_dataset\n",
    "    elif data_version == \"new\":\n",
    "        test_dataset = dataset[\"test\"]\n",
    "        return test_dataset\n",
    "    else:\n",
    "        raise ValueError(data_version)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "35ff7d5a-98f2-4360-8816-4143253fea8a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:33.864462Z",
     "iopub.status.busy": "2025-01-29T21:54:33.864299Z",
     "iopub.status.idle": "2025-01-29T21:54:33.868713Z",
     "shell.execute_reply": "2025-01-29T21:54:33.868350Z",
     "shell.execute_reply.started": "2025-01-29T21:54:33.864446Z"
    }
   },
   "outputs": [],
   "source": [
    "async def evaluate_text(\n",
    "    request: TextEvaluationRequest,\n",
    "    model_type: str = MODEL_TYPE,\n",
    "    # This should be an API query parameter, but it looks like the submission repo\n",
    "    # https://huggingface.co/spaces/frugal-ai-challenge/submission-portal\n",
    "    # is built in a way to not accept any other endpoints or parameters.\n",
    "):\n",
    "    \"\"\"\n",
    "    Evaluate text classification for climate disinformation detection.\n",
    "\n",
    "    Current Model: Random Baseline\n",
    "    - Makes random predictions from the label space (0-7)\n",
    "    - Used as a baseline for comparison\n",
    "    \"\"\"\n",
    "    # Get space info\n",
    "    username, space_url = get_space_info()\n",
    "\n",
    "    test_dataset = fetch_test_dataset(request, \"new\")\n",
    "\n",
    "    # Start tracking emissions\n",
    "    tracker.start()\n",
    "    tracker.start_task(\"inference\")\n",
    "\n",
    "    # --------------------------------------------------------------------------------------------\n",
    "    # YOUR MODEL INFERENCE CODE HERE\n",
    "    # Update the code below to replace the random baseline by your model inference within the inference pass where the energy consumption and emissions are tracked.\n",
    "    # --------------------------------------------------------------------------------------------\n",
    "\n",
    "    true_labels = test_dataset[\"label\"]\n",
    "    if model_type == \"baseline\":\n",
    "        predictions = baseline_model(len(true_labels))\n",
    "    elif model_type[:5] == \"bert-\":\n",
    "        predictions = bert_model(test_dataset, model_type)\n",
    "    else:\n",
    "        raise ValueError(model_type)\n",
    "\n",
    "    # --------------------------------------------------------------------------------------------\n",
    "    # YOUR MODEL INFERENCE STOPS HERE\n",
    "    # --------------------------------------------------------------------------------------------\n",
    "\n",
    "    # Stop tracking emissions\n",
    "    emissions_data = tracker.stop_task()\n",
    "\n",
    "    # Calculate accuracy\n",
    "    accuracy = accuracy_score(true_labels, predictions)\n",
    "\n",
    "    # Prepare results dictionary\n",
    "    results = {\n",
    "        \"username\": username,\n",
    "        \"space_url\": space_url,\n",
    "        \"submission_timestamp\": datetime.now().isoformat(),\n",
    "        \"model_description\": DESCRIPTIONS[model_type],\n",
    "        \"accuracy\": float(accuracy),\n",
    "        \"energy_consumed_wh\": emissions_data.energy_consumed * 1000,\n",
    "        \"emissions_gco2eq\": emissions_data.emissions * 1000,\n",
    "        \"emissions_data\": clean_emissions_data(emissions_data),\n",
    "        \"api_route\": ROUTE,\n",
    "        \"dataset_config\": {\n",
    "            \"dataset_name\": request.dataset_name,\n",
    "            \"test_size\": request.test_size,\n",
    "            \"test_seed\": request.test_seed,\n",
    "        },\n",
    "    }\n",
    "\n",
    "    return results, predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "de3404e5-45de-455b-a806-0ba511223ef4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:33.871157Z",
     "iopub.status.busy": "2025-01-29T21:54:33.870958Z",
     "iopub.status.idle": "2025-01-29T21:54:38.349553Z",
     "shell.execute_reply": "2025-01-29T21:54:38.349352Z",
     "shell.execute_reply.started": "2025-01-29T21:54:33.871143Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting my code block.\n",
      "Loading from model_repo: Nonnormalizable/frugal-ai-text-bert-mini\n",
      "Using device: mps\n",
      "Starting model run.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[codecarbon INFO @ 13:54:38] Energy consumed for RAM : 0.000005 kWh. RAM Power : 6.75 W\n",
      "[codecarbon INFO @ 13:54:38] Delta energy consumed for CPU with constant : 0.000029 kWh, power : 42.5 W\n",
      "[codecarbon INFO @ 13:54:38] Energy consumed for All CPU : 0.000029 kWh\n",
      "[codecarbon INFO @ 13:54:38] 0.000033 kWh of electricity used since the beginning.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "End of model run.\n",
      "End of my code block.\n"
     ]
    }
   ],
   "source": [
    "results, predictions = await evaluate_text(request=TextEvaluationRequest())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58b936f7-3326-4707-95f9-e7dcfd5f4129",
   "metadata": {},
   "source": [
    "With CPU on my laptop:\\\n",
    "`[codecarbon INFO @ 13:47:40] 0.000128 kWh of electricity used since the beginning.`\n",
    "\n",
    "With MPS on my laptop:\\\n",
    "`[codecarbon INFO @ 13:54:38] 0.000033 kWh of electricity used since the beginning.`"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "45b015dc-8afb-4537-84a6-0a7030647fa7",
   "metadata": {},
   "source": [
    "# Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a401c54-2d26-432e-bf7b-2d261d5b355e",
   "metadata": {},
   "source": [
    "## Construct dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "3d7b506f-9fc1-4ab5-91bc-4b732a021392",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:38.350066Z",
     "iopub.status.busy": "2025-01-29T21:54:38.349967Z",
     "iopub.status.idle": "2025-01-29T21:54:39.757646Z",
     "shell.execute_reply": "2025-01-29T21:54:39.757285Z",
     "shell.execute_reply.started": "2025-01-29T21:54:38.350056Z"
    }
   },
   "outputs": [],
   "source": [
    "test_dataset = fetch_test_dataset(TextEvaluationRequest(), \"new\")\n",
    "tokenizer = AutoTokenizer.from_pretrained(\n",
    "    f\"Nonnormalizable/frugal-ai-text-{MODEL_TYPE}\"\n",
    ")\n",
    "tokens = tokenizer(test_dataset[\"quote\"])\n",
    "token_lengths = [len(t) for t in tokens[\"input_ids\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e4af2a5b-704e-43f8-9fcb-e2f8b56b61ab",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.758142Z",
     "iopub.status.busy": "2025-01-29T21:54:39.758042Z",
     "iopub.status.idle": "2025-01-29T21:54:39.761545Z",
     "shell.execute_reply": "2025-01-29T21:54:39.761296Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.758132Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8580803937653815"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results[\"accuracy\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9ea094f1-cb30-4a69-a02e-734f0a0e0483",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.761989Z",
     "iopub.status.busy": "2025-01-29T21:54:39.761911Z",
     "iopub.status.idle": "2025-01-29T21:54:39.764092Z",
     "shell.execute_reply": "2025-01-29T21:54:39.763878Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.761980Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1219"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "9663b72e-51d1-41d6-8c30-768f6bbcba4d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.764484Z",
     "iopub.status.busy": "2025-01-29T21:54:39.764413Z",
     "iopub.status.idle": "2025-01-29T21:54:39.766356Z",
     "shell.execute_reply": "2025-01-29T21:54:39.766155Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.764476Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1046.0"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predictions) * results[\"accuracy\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7fb3dd9f-2278-40ff-b220-32dcda614662",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.766857Z",
     "iopub.status.busy": "2025-01-29T21:54:39.766753Z",
     "iopub.status.idle": "2025-01-29T21:54:39.768548Z",
     "shell.execute_reply": "2025-01-29T21:54:39.768342Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.766850Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6095.0"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predictions) / 0.20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c2c1d565-402f-48e0-beb2-70b9d3257945",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.768980Z",
     "iopub.status.busy": "2025-01-29T21:54:39.768909Z",
     "iopub.status.idle": "2025-01-29T21:54:39.787675Z",
     "shell.execute_reply": "2025-01-29T21:54:39.787330Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.768972Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "analysis_df = pd.DataFrame()\n",
    "for f in test_dataset.features:\n",
    "    c = pd.Series(test_dataset[f])\n",
    "    analysis_df[f] = c\n",
    "\n",
    "analysis_df[\"predicted_label\"] = pd.Series(predictions).astype(int)\n",
    "\n",
    "analysis_df = analysis_df.rename(\n",
    "    columns={\"__index_level_0__\": \"index_0\", \"label\": \"true_label\"}\n",
    ").set_index(\"index_0\", verify_integrity=True)\n",
    "analysis_df[\"quote_length_token\"] = token_lengths\n",
    "analysis_df = analysis_df.sort_index()\n",
    "analysis_df[\"is_correct\"] = (\n",
    "    analysis_df.true_label == analysis_df.predicted_label\n",
    ").astype(int)\n",
    "analysis_df = analysis_df.assign(is_wrong=1 - analysis_df.is_correct)\n",
    "\n",
    "analysis_df[\"quote_length_char\"] = analysis_df[\"quote\"].str.len()\n",
    "analysis_df[\"quote_length_word\"] = analysis_df[\"quote\"].str.split().apply(len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4edfa8af-ca6c-48f9-9d41-3856b2fb4a4e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.788243Z",
     "iopub.status.busy": "2025-01-29T21:54:39.788111Z",
     "iopub.status.idle": "2025-01-29T21:54:39.794680Z",
     "shell.execute_reply": "2025-01-29T21:54:39.794446Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.788233Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>quote</th>\n",
       "      <th>true_label</th>\n",
       "      <th>source</th>\n",
       "      <th>url</th>\n",
       "      <th>language</th>\n",
       "      <th>subsource</th>\n",
       "      <th>id</th>\n",
       "      <th>predicted_label</th>\n",
       "      <th>quote_length_token</th>\n",
       "      <th>is_correct</th>\n",
       "      <th>is_wrong</th>\n",
       "      <th>quote_length_char</th>\n",
       "      <th>quote_length_word</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>index_0</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5011</th>\n",
       "      <td>It’s very costly to operate here in America co...</td>\n",
       "      <td>0</td>\n",
       "      <td>Desmog</td>\n",
       "      <td>https://www.desmog.com/harold-hamm/</td>\n",
       "      <td>en</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "      <td>68</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>311</td>\n",
       "      <td>55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3523</th>\n",
       "      <td>It would be necessary to use ambiguous and mis...</td>\n",
       "      <td>0</td>\n",
       "      <td>Desmog</td>\n",
       "      <td>https://www.desmog.com/george-pearson/</td>\n",
       "      <td>en</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "      <td>47</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>238</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3641</th>\n",
       "      <td>Unfortunately, Canada doesn’t have a Nigel Law...</td>\n",
       "      <td>4</td>\n",
       "      <td>Desmog</td>\n",
       "      <td>https://www.desmog.com/canadians-for-affordabl...</td>\n",
       "      <td>en</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>4</td>\n",
       "      <td>122</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>572</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>746</th>\n",
       "      <td>This is recognised by the major carbon dioxide...</td>\n",
       "      <td>0</td>\n",
       "      <td>FLICC</td>\n",
       "      <td>https://huggingface.co/datasets/fzanartu/FLICC...</td>\n",
       "      <td>en</td>\n",
       "      <td>jintrain</td>\n",
       "      <td>None</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>140</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2273</th>\n",
       "      <td>Climate change is a genuine problem that will ...</td>\n",
       "      <td>0</td>\n",
       "      <td>Desmog</td>\n",
       "      <td>https://www.desmog.com/bjorn-lomborg/</td>\n",
       "      <td>en</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>3</td>\n",
       "      <td>102</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>510</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     quote  true_label  \\\n",
       "index_0                                                                  \n",
       "5011     It’s very costly to operate here in America co...           0   \n",
       "3523     It would be necessary to use ambiguous and mis...           0   \n",
       "3641     Unfortunately, Canada doesn’t have a Nigel Law...           4   \n",
       "746      This is recognised by the major carbon dioxide...           0   \n",
       "2273     Climate change is a genuine problem that will ...           0   \n",
       "\n",
       "         source                                                url language  \\\n",
       "index_0                                                                       \n",
       "5011     Desmog                https://www.desmog.com/harold-hamm/       en   \n",
       "3523     Desmog             https://www.desmog.com/george-pearson/       en   \n",
       "3641     Desmog  https://www.desmog.com/canadians-for-affordabl...       en   \n",
       "746       FLICC  https://huggingface.co/datasets/fzanartu/FLICC...       en   \n",
       "2273     Desmog              https://www.desmog.com/bjorn-lomborg/       en   \n",
       "\n",
       "        subsource    id  predicted_label  quote_length_token  is_correct  \\\n",
       "index_0                                                                    \n",
       "5011         None  None                0                  68           1   \n",
       "3523         None  None                0                  47           1   \n",
       "3641         None  None                4                 122           1   \n",
       "746      jintrain  None                0                  35           1   \n",
       "2273         None  None                3                 102           0   \n",
       "\n",
       "         is_wrong  quote_length_char  quote_length_word  \n",
       "index_0                                                  \n",
       "5011            0                311                 55  \n",
       "3523            0                238                 40  \n",
       "3641            0                572                 98  \n",
       "746             0                140                 26  \n",
       "2273            1                510                 79  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "cbc638e3-312c-4f5e-8abc-dd07df4c8749",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.795196Z",
     "iopub.status.busy": "2025-01-29T21:54:39.795100Z",
     "iopub.status.idle": "2025-01-29T21:54:39.803187Z",
     "shell.execute_reply": "2025-01-29T21:54:39.802949Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.795187Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>true_label</th>\n",
       "      <th>predicted_label</th>\n",
       "      <th>quote_length_token</th>\n",
       "      <th>is_correct</th>\n",
       "      <th>is_wrong</th>\n",
       "      <th>quote_length_char</th>\n",
       "      <th>quote_length_word</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1219.000000</td>\n",
       "      <td>1219.000000</td>\n",
       "      <td>1219.000000</td>\n",
       "      <td>1219.000000</td>\n",
       "      <td>1219.000000</td>\n",
       "      <td>1219.000000</td>\n",
       "      <td>1219.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2.828548</td>\n",
       "      <td>2.908942</td>\n",
       "      <td>60.122231</td>\n",
       "      <td>0.858080</td>\n",
       "      <td>0.141920</td>\n",
       "      <td>279.809680</td>\n",
       "      <td>46.735029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2.322724</td>\n",
       "      <td>2.278062</td>\n",
       "      <td>46.903184</td>\n",
       "      <td>0.349111</td>\n",
       "      <td>0.349111</td>\n",
       "      <td>226.966534</td>\n",
       "      <td>37.764322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>136.000000</td>\n",
       "      <td>23.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>222.000000</td>\n",
       "      <td>37.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>77.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>360.500000</td>\n",
       "      <td>60.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>561.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2643.000000</td>\n",
       "      <td>454.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        true_label  predicted_label  quote_length_token   is_correct  \\\n",
       "count  1219.000000      1219.000000         1219.000000  1219.000000   \n",
       "mean      2.828548         2.908942           60.122231     0.858080   \n",
       "std       2.322724         2.278062           46.903184     0.349111   \n",
       "min       0.000000         0.000000            7.000000     0.000000   \n",
       "25%       0.000000         1.000000           30.000000     1.000000   \n",
       "50%       3.000000         3.000000           48.000000     1.000000   \n",
       "75%       5.000000         5.000000           77.000000     1.000000   \n",
       "max       7.000000         7.000000          561.000000     1.000000   \n",
       "\n",
       "          is_wrong  quote_length_char  quote_length_word  \n",
       "count  1219.000000        1219.000000        1219.000000  \n",
       "mean      0.141920         279.809680          46.735029  \n",
       "std       0.349111         226.966534          37.764322  \n",
       "min       0.000000          24.000000           4.000000  \n",
       "25%       0.000000         136.000000          23.000000  \n",
       "50%       0.000000         222.000000          37.000000  \n",
       "75%       0.000000         360.500000          60.000000  \n",
       "max       1.000000        2643.000000         454.000000  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "29f1a80e-abc5-4c51-8068-f02cb2f8958b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.803672Z",
     "iopub.status.busy": "2025-01-29T21:54:39.803586Z",
     "iopub.status.idle": "2025-01-29T21:54:39.807223Z",
     "shell.execute_reply": "2025-01-29T21:54:39.807004Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.803664Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "quote                 0.000000\n",
       "true_label            0.000000\n",
       "source                0.000000\n",
       "url                   0.000000\n",
       "language              0.000000\n",
       "subsource             0.675144\n",
       "id                    1.000000\n",
       "predicted_label       0.000000\n",
       "quote_length_token    0.000000\n",
       "is_correct            0.000000\n",
       "is_wrong              0.000000\n",
       "quote_length_char     0.000000\n",
       "quote_length_word     0.000000\n",
       "dtype: float64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.isna().mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "9536ddbf-8780-4091-a161-3926003968d4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.807686Z",
     "iopub.status.busy": "2025-01-29T21:54:39.807594Z",
     "iopub.status.idle": "2025-01-29T21:54:39.811445Z",
     "shell.execute_reply": "2025-01-29T21:54:39.811226Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.807677Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>quote_length_char</th>\n",
       "      <th>quote_length_word</th>\n",
       "      <th>quote_length_token</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>quote_length_char</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.992111</td>\n",
       "      <td>0.988758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>quote_length_word</th>\n",
       "      <td>0.992111</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.993291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>quote_length_token</th>\n",
       "      <td>0.988758</td>\n",
       "      <td>0.993291</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    quote_length_char  quote_length_word  quote_length_token\n",
       "quote_length_char            1.000000           0.992111            0.988758\n",
       "quote_length_word            0.992111           1.000000            0.993291\n",
       "quote_length_token           0.988758           0.993291            1.000000"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df[[\"quote_length_char\", \"quote_length_word\", \"quote_length_token\"]].corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7cf801d-9c25-4d46-af1d-17dad1422251",
   "metadata": {},
   "source": [
    "## Error distribution across classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "d88ee6e0-d7fd-4f11-8c09-5f118913a88b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.811949Z",
     "iopub.status.busy": "2025-01-29T21:54:39.811855Z",
     "iopub.status.idle": "2025-01-29T21:54:39.820877Z",
     "shell.execute_reply": "2025-01-29T21:54:39.820601Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.811941Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>true_label</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>predicted_label</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>252.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.0</td>\n",
       "      <td>143.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>125.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>136.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>13.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>136.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>16.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>123.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>51.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "true_label           0      1      2     3      4      5      6     7\n",
       "predicted_label                                                      \n",
       "0                252.0    4.0    1.0   2.0    6.0    0.0    0.0   1.0\n",
       "1                  3.0  143.0    3.0   7.0    1.0    6.0    3.0   0.0\n",
       "2                  5.0    4.0  125.0   5.0    0.0   15.0    1.0   3.0\n",
       "3                  8.0    0.0    3.0  80.0    3.0    0.0    0.0   0.0\n",
       "4                  9.0    0.0    1.0   1.0  136.0    0.0    3.0  10.0\n",
       "5                 13.0    1.0    4.0   1.0    7.0  136.0    6.0   0.0\n",
       "6                 16.0    2.0    0.0   1.0    2.0    3.0  123.0   0.0\n",
       "7                  1.0    0.0    0.0   0.0    5.0    0.0    3.0  51.0"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "heatmap_df = analysis_df.pivot_table(\n",
    "    columns=\"true_label\", index=\"predicted_label\", aggfunc=\"count\", values=\"quote\"\n",
    ").fillna(0)\n",
    "heatmap_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "c95fc4d5-e217-4ee8-b4f8-9018a2a6aa88",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.821448Z",
     "iopub.status.busy": "2025-01-29T21:54:39.821321Z",
     "iopub.status.idle": "2025-01-29T21:54:39.915804Z",
     "shell.execute_reply": "2025-01-29T21:54:39.915543Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.821438Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='true_label', ylabel='predicted_label'>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 435,
       "width": 540
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(heatmap_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "12d2177e-bae5-4b45-b47d-b8ee8efb2112",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.916296Z",
     "iopub.status.busy": "2025-01-29T21:54:39.916221Z",
     "iopub.status.idle": "2025-01-29T21:54:39.989195Z",
     "shell.execute_reply": "2025-01-29T21:54:39.988907Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.916288Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='true_label', ylabel='predicted_label'>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 433,
       "width": 535
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(np.log10(heatmap_df + 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "4452036d-52f3-4390-80d1-0f4fbf3db9d6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.989708Z",
     "iopub.status.busy": "2025-01-29T21:54:39.989634Z",
     "iopub.status.idle": "2025-01-29T21:54:39.993744Z",
     "shell.execute_reply": "2025-01-29T21:54:39.993412Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.989700Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_wrong</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>true_label</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.215385</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.179153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.175258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.150000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.150000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.115108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.087591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.071429</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            is_wrong\n",
       "true_label          \n",
       "7           0.215385\n",
       "0           0.179153\n",
       "3           0.175258\n",
       "4           0.150000\n",
       "5           0.150000\n",
       "6           0.115108\n",
       "2           0.087591\n",
       "1           0.071429"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby(\"true_label\").agg({\"is_wrong\": \"mean\"}).sort_values(\n",
    "    \"is_wrong\", ascending=False\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "3b684b9c-ab60-406d-a65f-2b31ceb36a4f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.994197Z",
     "iopub.status.busy": "2025-01-29T21:54:39.994123Z",
     "iopub.status.idle": "2025-01-29T21:54:39.998021Z",
     "shell.execute_reply": "2025-01-29T21:54:39.997782Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.994189Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_wrong</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>true_label</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            is_wrong\n",
       "true_label          \n",
       "0                 55\n",
       "4                 24\n",
       "5                 24\n",
       "3                 17\n",
       "6                 16\n",
       "7                 14\n",
       "2                 12\n",
       "1                 11"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby(\"true_label\").agg({\"is_wrong\": \"sum\"}).sort_values(\n",
    "    \"is_wrong\", ascending=False\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cf7cd28-6ef3-4fb3-9744-86a1d5daaa5f",
   "metadata": {},
   "source": [
    "## Quote length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "34f5373b-f382-40f3-b6a3-f4ba4885c54b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:39.998564Z",
     "iopub.status.busy": "2025-01-29T21:54:39.998436Z",
     "iopub.status.idle": "2025-01-29T21:54:40.007698Z",
     "shell.execute_reply": "2025-01-29T21:54:40.007466Z",
     "shell.execute_reply.started": "2025-01-29T21:54:39.998554Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"8\" valign=\"top\">quote_length_char</th>\n",
       "      <th>count</th>\n",
       "      <td>173.0</td>\n",
       "      <td>1046.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>298.0</td>\n",
       "      <td>276.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>262.0</td>\n",
       "      <td>220.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>34.0</td>\n",
       "      <td>24.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>129.0</td>\n",
       "      <td>136.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>221.0</td>\n",
       "      <td>224.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>368.0</td>\n",
       "      <td>360.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1546.0</td>\n",
       "      <td>2643.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"8\" valign=\"top\">quote_length_word</th>\n",
       "      <th>count</th>\n",
       "      <td>173.0</td>\n",
       "      <td>1046.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>49.6</td>\n",
       "      <td>46.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>42.7</td>\n",
       "      <td>36.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>22.0</td>\n",
       "      <td>23.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>37.0</td>\n",
       "      <td>37.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>60.0</td>\n",
       "      <td>59.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>273.0</td>\n",
       "      <td>454.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"8\" valign=\"top\">quote_length_token</th>\n",
       "      <th>count</th>\n",
       "      <td>173.0</td>\n",
       "      <td>1046.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>63.2</td>\n",
       "      <td>59.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>51.8</td>\n",
       "      <td>46.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>11.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>30.0</td>\n",
       "      <td>31.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>49.0</td>\n",
       "      <td>48.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>77.0</td>\n",
       "      <td>77.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>337.0</td>\n",
       "      <td>561.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "is_correct                     0       1\n",
       "quote_length_char  count   173.0  1046.0\n",
       "                   mean    298.0   276.8\n",
       "                   std     262.0   220.6\n",
       "                   min      34.0    24.0\n",
       "                   25%     129.0   136.0\n",
       "                   50%     221.0   224.0\n",
       "                   75%     368.0   360.0\n",
       "                   max    1546.0  2643.0\n",
       "quote_length_word  count   173.0  1046.0\n",
       "                   mean     49.6    46.3\n",
       "                   std      42.7    36.9\n",
       "                   min       7.0     4.0\n",
       "                   25%      22.0    23.0\n",
       "                   50%      37.0    37.0\n",
       "                   75%      60.0    59.0\n",
       "                   max     273.0   454.0\n",
       "quote_length_token count   173.0  1046.0\n",
       "                   mean     63.2    59.6\n",
       "                   std      51.8    46.1\n",
       "                   min      11.0     7.0\n",
       "                   25%      30.0    31.0\n",
       "                   50%      49.0    48.0\n",
       "                   75%      77.0    77.0\n",
       "                   max     337.0   561.0"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby(\"is_correct\")[\n",
    "    [\"quote_length_char\", \"quote_length_word\", \"quote_length_token\"]\n",
    "].describe().transpose().round(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "5891626a-f40c-43dc-9f1b-e5146dcb5665",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.008167Z",
     "iopub.status.busy": "2025-01-29T21:54:40.008080Z",
     "iopub.status.idle": "2025-01-29T21:54:40.016873Z",
     "shell.execute_reply": "2025-01-29T21:54:40.016654Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.008159Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>source</th>\n",
       "      <th>Desmog</th>\n",
       "      <th>FLICC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"8\" valign=\"top\">quote_length_char</th>\n",
       "      <th>count</th>\n",
       "      <td>823.0</td>\n",
       "      <td>396.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>321.9</td>\n",
       "      <td>192.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>251.6</td>\n",
       "      <td>125.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>44.0</td>\n",
       "      <td>24.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>155.0</td>\n",
       "      <td>106.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>259.0</td>\n",
       "      <td>159.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>401.5</td>\n",
       "      <td>248.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2643.0</td>\n",
       "      <td>878.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"8\" valign=\"top\">quote_length_word</th>\n",
       "      <th>count</th>\n",
       "      <td>823.0</td>\n",
       "      <td>396.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>53.7</td>\n",
       "      <td>32.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>41.9</td>\n",
       "      <td>20.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>11.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>26.5</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>43.0</td>\n",
       "      <td>27.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>67.0</td>\n",
       "      <td>42.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>454.0</td>\n",
       "      <td>140.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"8\" valign=\"top\">quote_length_token</th>\n",
       "      <th>count</th>\n",
       "      <td>823.0</td>\n",
       "      <td>396.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>68.7</td>\n",
       "      <td>42.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>52.1</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>14.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>35.0</td>\n",
       "      <td>25.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>56.0</td>\n",
       "      <td>36.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>85.0</td>\n",
       "      <td>54.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>561.0</td>\n",
       "      <td>176.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "source                    Desmog  FLICC\n",
       "quote_length_char  count   823.0  396.0\n",
       "                   mean    321.9  192.3\n",
       "                   std     251.6  125.5\n",
       "                   min      44.0   24.0\n",
       "                   25%     155.0  106.0\n",
       "                   50%     259.0  159.0\n",
       "                   75%     401.5  248.5\n",
       "                   max    2643.0  878.0\n",
       "quote_length_word  count   823.0  396.0\n",
       "                   mean     53.7   32.2\n",
       "                   std      41.9   20.6\n",
       "                   min      11.0    4.0\n",
       "                   25%      26.5   18.0\n",
       "                   50%      43.0   27.0\n",
       "                   75%      67.0   42.0\n",
       "                   max     454.0  140.0\n",
       "quote_length_token count   823.0  396.0\n",
       "                   mean     68.7   42.3\n",
       "                   std      52.1   26.0\n",
       "                   min      14.0    7.0\n",
       "                   25%      35.0   25.0\n",
       "                   50%      56.0   36.0\n",
       "                   75%      85.0   54.0\n",
       "                   max     561.0  176.0"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby([\"source\"])[\n",
    "    [\"quote_length_char\", \"quote_length_word\", \"quote_length_token\"]\n",
    "].describe().transpose().round(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4808214d-c956-49ec-834b-efbad9f06eaa",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.017365Z",
     "iopub.status.busy": "2025-01-29T21:54:40.017242Z",
     "iopub.status.idle": "2025-01-29T21:54:40.022514Z",
     "shell.execute_reply": "2025-01-29T21:54:40.022272Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.017356Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>true_label</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>quote_length_char</th>\n",
       "      <td>267.5</td>\n",
       "      <td>195.4</td>\n",
       "      <td>285.9</td>\n",
       "      <td>258.0</td>\n",
       "      <td>342.8</td>\n",
       "      <td>308.9</td>\n",
       "      <td>278.3</td>\n",
       "      <td>334.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>quote_length_word</th>\n",
       "      <td>44.6</td>\n",
       "      <td>34.0</td>\n",
       "      <td>47.9</td>\n",
       "      <td>43.3</td>\n",
       "      <td>56.4</td>\n",
       "      <td>51.6</td>\n",
       "      <td>45.9</td>\n",
       "      <td>55.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>quote_length_token</th>\n",
       "      <td>57.8</td>\n",
       "      <td>44.0</td>\n",
       "      <td>60.9</td>\n",
       "      <td>55.9</td>\n",
       "      <td>71.6</td>\n",
       "      <td>66.4</td>\n",
       "      <td>60.1</td>\n",
       "      <td>70.6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "true_label              0      1      2      3      4      5      6      7\n",
       "quote_length_char   267.5  195.4  285.9  258.0  342.8  308.9  278.3  334.4\n",
       "quote_length_word    44.6   34.0   47.9   43.3   56.4   51.6   45.9   55.8\n",
       "quote_length_token   57.8   44.0   60.9   55.9   71.6   66.4   60.1   70.6"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby([\"true_label\"])[\n",
    "    [\"quote_length_char\", \"quote_length_word\", \"quote_length_token\"]\n",
    "].mean().transpose().round(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3f9d3d01-9bf6-430b-b13b-aa52a17d4648",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "8ba0f8ab-cb44-494c-be73-2d396340286a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.022981Z",
     "iopub.status.busy": "2025-01-29T21:54:40.022882Z",
     "iopub.status.idle": "2025-01-29T21:54:40.259116Z",
     "shell.execute_reply": "2025-01-29T21:54:40.258810Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.022974Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/6x/4bcqvs9d4cvcnt_qj6jpmj300000gn/T/ipykernel_78254/1952994841.py:3: UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator.\n",
      "  labels = ax.set_xticklabels(ax.get_xticklabels(), rotation=60)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 519,
       "width": 599
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "analysis_df[\"token_len_q\"] = pd.qcut(analysis_df[\"quote_length_token\"], q=25)\n",
    "ax = sns.pointplot(analysis_df, x=\"token_len_q\", y=\"is_correct\")\n",
    "labels = ax.set_xticklabels(ax.get_xticklabels(), rotation=60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "6066b3c1-d01f-48ed-b5d3-a29dd8e7517f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.259594Z",
     "iopub.status.busy": "2025-01-29T21:54:40.259516Z",
     "iopub.status.idle": "2025-01-29T21:54:40.264283Z",
     "shell.execute_reply": "2025-01-29T21:54:40.264017Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.259586Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>token_len_q_manual</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.862393</td>\n",
       "      <td>1170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.755102</td>\n",
       "      <td>49</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    is_correct  quote\n",
       "token_len_q_manual                   \n",
       "0                     0.862393   1170\n",
       "1                     0.755102     49"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df[\"token_len_q_manual\"] = (analysis_df[\"quote_length_token\"] > 145).astype(\n",
    "    int\n",
    ")\n",
    "analysis_df.groupby(\"token_len_q_manual\").agg({\"is_correct\": \"mean\", \"quote\": \"count\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "e170c7c7-019f-4ebd-b9c1-503dd9aa516e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.264734Z",
     "iopub.status.busy": "2025-01-29T21:54:40.264660Z",
     "iopub.status.idle": "2025-01-29T21:54:40.268523Z",
     "shell.execute_reply": "2025-01-29T21:54:40.268278Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.264726Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>token_len_q_manual</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.858091</td>\n",
       "      <td>1205</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.857143</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    is_correct  quote\n",
       "token_len_q_manual                   \n",
       "0                     0.858091   1205\n",
       "1                     0.857143     14"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df[\"token_len_q_manual\"] = (analysis_df[\"quote_length_token\"] > 256).astype(\n",
    "    int\n",
    ")\n",
    "analysis_df.groupby(\"token_len_q_manual\").agg({\"is_correct\": \"mean\", \"quote\": \"count\"})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "def27ad7-7c99-4a5e-8139-910d324e1068",
   "metadata": {},
   "source": [
    "## Source, etc."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "bc30d7f8-b37c-4dec-b60b-05d1d40cb2d5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.269126Z",
     "iopub.status.busy": "2025-01-29T21:54:40.268954Z",
     "iopub.status.idle": "2025-01-29T21:54:40.273158Z",
     "shell.execute_reply": "2025-01-29T21:54:40.272942Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.269116Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Desmog</th>\n",
       "      <td>0.851762</td>\n",
       "      <td>823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLICC</th>\n",
       "      <td>0.871212</td>\n",
       "      <td>396</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        is_correct  quote\n",
       "source                   \n",
       "Desmog    0.851762    823\n",
       "FLICC     0.871212    396"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby([\"source\"], dropna=False).agg(\n",
    "    {\"is_correct\": \"mean\", \"quote\": \"count\"}\n",
    ").sort_values(\"quote\", ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "7551b1f7-6e9c-4623-89eb-fdf4de411b38",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.273616Z",
     "iopub.status.busy": "2025-01-29T21:54:40.273527Z",
     "iopub.status.idle": "2025-01-29T21:54:40.278307Z",
     "shell.execute_reply": "2025-01-29T21:54:40.278073Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.273608Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <th>subsource</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Desmog</th>\n",
       "      <th>NaN</th>\n",
       "      <td>0.851762</td>\n",
       "      <td>823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"10\" valign=\"top\">FLICC</th>\n",
       "      <th>CARDS</th>\n",
       "      <td>0.900621</td>\n",
       "      <td>161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hamburg_test3</th>\n",
       "      <td>0.829545</td>\n",
       "      <td>88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jintrain</th>\n",
       "      <td>0.947368</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hamburg_test2</th>\n",
       "      <td>0.766667</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Alhindi_train</th>\n",
       "      <td>0.875000</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hamburg_test1</th>\n",
       "      <td>0.833333</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jintest</th>\n",
       "      <td>0.818182</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Alhindi_dev</th>\n",
       "      <td>0.800000</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jindev</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Alhindi_test</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      is_correct  quote\n",
       "source subsource                       \n",
       "Desmog NaN              0.851762    823\n",
       "FLICC  CARDS            0.900621    161\n",
       "       hamburg_test3    0.829545     88\n",
       "       jintrain         0.947368     38\n",
       "       hamburg_test2    0.766667     30\n",
       "       Alhindi_train    0.875000     24\n",
       "       hamburg_test1    0.833333     24\n",
       "       jintest          0.818182     11\n",
       "       Alhindi_dev      0.800000     10\n",
       "       jindev           1.000000      9\n",
       "       Alhindi_test     1.000000      1"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby([\"source\", \"subsource\"], dropna=False).agg(\n",
    "    {\"is_correct\": \"mean\", \"quote\": \"count\"}\n",
    ").sort_values(\"quote\", ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "6e1ad460-28bc-48cb-9f29-9554bedf1bc2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.278880Z",
     "iopub.status.busy": "2025-01-29T21:54:40.278698Z",
     "iopub.status.idle": "2025-01-29T21:54:40.282717Z",
     "shell.execute_reply": "2025-01-29T21:54:40.282486Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.278871Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>language</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>en</th>\n",
       "      <td>0.85808</td>\n",
       "      <td>1219</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          is_correct  quote\n",
       "language                   \n",
       "en           0.85808   1219"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby([\"language\"], dropna=False).agg(\n",
    "    {\"is_correct\": \"mean\", \"quote\": \"count\"}\n",
    ").sort_values(\"quote\", ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "eeba5d29-7888-4456-95fd-3e513c3bb17a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.283158Z",
     "iopub.status.busy": "2025-01-29T21:54:40.283085Z",
     "iopub.status.idle": "2025-01-29T21:54:40.287463Z",
     "shell.execute_reply": "2025-01-29T21:54:40.287252Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.283151Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>is_correct</th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>url</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>https://huggingface.co/datasets/fzanartu/FLICCdataset</th>\n",
       "      <td>0.871212</td>\n",
       "      <td>396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/manning-foundation-for-democratic-education/</th>\n",
       "      <td>0.800000</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/marlo-lewis-jr/</th>\n",
       "      <td>0.700000</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/lee-raymond/</th>\n",
       "      <td>0.600000</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/michael-shellenberger/</th>\n",
       "      <td>0.555556</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/bjorn-lomborg/</th>\n",
       "      <td>0.625000</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/fred-palmer/</th>\n",
       "      <td>0.714286</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/naomi-seibt/</th>\n",
       "      <td>0.833333</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/myron-ebell/</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>https://www.desmog.com/william-briggs/</th>\n",
       "      <td>0.833333</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    is_correct  quote\n",
       "url                                                                  \n",
       "https://huggingface.co/datasets/fzanartu/FLICCd...    0.871212    396\n",
       "https://www.desmog.com/manning-foundation-for-d...    0.800000     10\n",
       "https://www.desmog.com/marlo-lewis-jr/                0.700000     10\n",
       "https://www.desmog.com/lee-raymond/                   0.600000     10\n",
       "https://www.desmog.com/michael-shellenberger/         0.555556      9\n",
       "https://www.desmog.com/bjorn-lomborg/                 0.625000      8\n",
       "https://www.desmog.com/fred-palmer/                   0.714286      7\n",
       "https://www.desmog.com/naomi-seibt/                   0.833333      6\n",
       "https://www.desmog.com/myron-ebell/                   1.000000      6\n",
       "https://www.desmog.com/william-briggs/                0.833333      6"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.groupby([\"url\"], dropna=False).agg(\n",
    "    {\"is_correct\": \"mean\", \"quote\": \"count\"}\n",
    ").sort_values(\"quote\", ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5dd5e7df-3ea0-4d70-bbd1-6f9994c5ab82",
   "metadata": {},
   "source": [
    "## Look at examples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "3eed1379-1986-4d85-8c95-3a8d83c565fb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.287878Z",
     "iopub.status.busy": "2025-01-29T21:54:40.287810Z",
     "iopub.status.idle": "2025-01-29T21:54:40.289846Z",
     "shell.execute_reply": "2025-01-29T21:54:40.289629Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.287871Z"
    }
   },
   "outputs": [],
   "source": [
    "def print_example(i, row, errors_categorized=None):\n",
    "    print(\n",
    "        i,\n",
    "        row.source,\n",
    "        row.subsource,\n",
    "        f\"true={LABEL_MAPPING_INV[row.true_label]}\",\n",
    "        f\"predict={LABEL_MAPPING_INV[row.predicted_label]}\",\n",
    "    )\n",
    "    print(textwrap.fill(row.quote, width=80))\n",
    "    if errors_categorized is not None:\n",
    "        print(f\"My categorization: {errors_categorized[i]}\")\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "57dbb0ee-52a6-4166-a054-b68896641bcd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.290233Z",
     "iopub.status.busy": "2025-01-29T21:54:40.290168Z",
     "iopub.status.idle": "2025-01-29T21:54:40.296112Z",
     "shell.execute_reply": "2025-01-29T21:54:40.295845Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.290226Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "198 FLICC CARDS true=5_science_unreliable predict=5_science_unreliable\n",
      "The inference is that climate model predictions should be used to drive global\n",
      "energy policy. This is the simple linear model of scientism, that has been\n",
      "resoundingly debunked particularly for a complex problem like climate change.\n",
      "\n",
      "5057 Desmog None true=0_not_relevant predict=0_not_relevant\n",
      "So you don’t think we’re headed to two degrees on our current path?\n",
      "\n",
      "3372 Desmog None true=6_proponents_biased predict=6_proponents_biased\n",
      "Schemes and fibs by environmental extremists like Al Gore, Tom Steyer, and\n",
      "leaders at Google are killing jobs, increasing the price of energy, devouring\n",
      "tax dollars, and threatening America’s energy independence — all in an effort to\n",
      "make a quick buck.\n",
      "\n",
      "681 FLICC CARDS true=5_science_unreliable predict=5_science_unreliable\n",
      "First up, a common misconception in the climate debate is that radiation from a\n",
      "cold body (e.g. the -18C atmosphere) can warm a hot body (e.g. the +15C Earth\n",
      "surface) just because the cold body does indeed send very-low-energy photons to\n",
      "the hot body. Heat transfer (not radiation) from cold to hot is forbidden by the\n",
      "2nd Law of Thermodynamics on a macro basis, and by the Pauli Exclusion Principle\n",
      "of fundamental quantum theory on an atomic and molecular basis.\n",
      "\n",
      "1488 FLICC hamburg_test1 true=6_proponents_biased predict=6_proponents_biased\n",
      "Nothing like Markle who preaches about the global warming and used a private jet\n",
      "11 times in 2 weeks\n",
      "\n",
      "426 FLICC CARDS true=1_not_happening predict=1_not_happening\n",
      "These tide gauges show that sea levels are rising and falling around Vanuatu\n",
      "over the last 20 years (feast your eyes, there is a 30cm range on that graph\n",
      "below). Where is that CO2 signal? Seas around Vanuatu have been falling since\n",
      "2008.\n",
      "\n",
      "1056 FLICC CARDS true=1_not_happening predict=1_not_happening\n",
      "The world has not warmed up very much since the millennium. Twelve years is a\n",
      "reasonable time it (the temperature) has stayed almost constant, whereas it\n",
      "should have been rising carbon dioxide is rising, no question about that.\n",
      "\n",
      "1569 FLICC Alhindi_dev true=2_not_human predict=2_not_human\n",
      "at the end of recent ice ages, the concentration of carbon dioxide in the\n",
      "atmosphere started to rise only after temperatures began to climb.\n",
      "\n",
      "1562 FLICC hamburg_test3 true=5_science_unreliable predict=5_science_unreliable\n",
      "mann fixed his hockey stick numbers\n",
      "\n",
      "4853 Desmog None true=1_not_happening predict=1_not_happening\n",
      "Climate change – Sun & the stars vs C02 – I,\n",
      "\n",
      "3773 Desmog None true=5_science_unreliable predict=5_science_unreliable\n",
      "The difference in the global warming scare and the  crash that started in 1998\n",
      "was that  had a definite end date which proved it to be a marketing ploy by\n",
      "software companies to sell unnecessary updates to data backup systems whereas\n",
      "the climate change hype could go on indefinitely and that is what allows it to\n",
      "build and fester. … e’re constantly being bombarded with rhetoric that 97% of\n",
      "all scientists believe that mankind has played a role in changing the earth’s\n",
      "climate, but common sense tells us no large group of people on our planet could\n",
      "ever reach 97% agreement on anything, even the world being round!\n",
      "\n",
      "3830 Desmog None true=0_not_relevant predict=0_not_relevant\n",
      "We’ve already got 250 of them and we’ve got 10,000 football-pitches’ worth of\n",
      "new solar farms in the pipeline.\n",
      "\n",
      "485 FLICC Alhindi_dev true=1_not_happening predict=1_not_happening\n",
      "there has been no systematic increase in the frequency of extreme weather\n",
      "events,\n",
      "\n",
      "1170 FLICC CARDS true=6_proponents_biased predict=6_proponents_biased\n",
      "This morning I read a terrific recounting of the the fact that a whole section\n",
      "of the scientific establishment is defending Gleick on the grounds that it's OK\n",
      "to lie to promote their cause.\n",
      "\n",
      "476 FLICC hamburg_test2 true=1_not_happening predict=1_not_happening\n",
      "Sea levels are not and will not rise. Show me and others the actual data that\n",
      "shows that they have in the last hundred years. Climate change? Climate\n",
      "bollocks.\n",
      "\n",
      "6018 Desmog None true=5_science_unreliable predict=5_science_unreliable\n",
      "We worry the sole focus on greenhouse gases and the unwise reliance on imperfect\n",
      "climate models while ignoring real data may leave civilization unprepared for a\n",
      "sudden climate shift that history tells us will occur again, very possibly soon.\n",
      "\n",
      "1599 FLICC jintrain true=6_proponents_biased predict=6_proponents_biased\n",
      "This debate as I argue at some length in Watermelons was always about left-wing\n",
      "ideology, quasi-religious hysteria, and follow the money corruption, never about\n",
      "\"science.\"\n",
      "\n",
      "2815 Desmog None true=0_not_relevant predict=0_not_relevant\n",
      "Public unease about safety and problems with costs, liability, and permanent\n",
      "storage do not make a flourishing nuclear industry impossible, but they do\n",
      "demonstrate the enormous influence that mistaken public risk perception can have\n",
      "on government policy and reveal the consistently inept bureaucratic handling of\n",
      "the challenge so far,” Smil wrote at the American Enterprise Institute‘s\n",
      "blog, AEIdeas.1 uclear energy’s discouraging record is even more unfortunate\n",
      "given that nuclear generation is the only low-carbon-footprint energy option\n",
      "readily available on a gigawatt-level scale. This is why nuclear power should be\n",
      "part of any serious attempt to reduce the rate of global warming. At the same\n",
      "time, it would be naïve to think that nuclear power could be (as some suggest)\n",
      "the single most effective tool for combating climate change in the next ten to\n",
      "30 years. The best hope is for it to offer a modest contribution.\n",
      "\n",
      "1042 FLICC jintrain true=1_not_happening predict=1_not_happening\n",
      "Furthermore, whereas in 2008 most of the ice was extremely thin, this year most\n",
      "has been at least two metres thick.\n",
      "\n",
      "3388 Desmog None true=0_not_relevant predict=0_not_relevant\n",
      "The  government needs to stop meddling in industries and create an atmosphere\n",
      "that allows business to prosper without pledging taxpayer support.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i, row in (\n",
    "    analysis_df.query(\"is_correct == 1\").sample(20, random_state=1).iterrows()\n",
    "):\n",
    "    print_example(i, row)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "5f59e1ed-7e62-4078-8bd6-199b95131dc4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.296751Z",
     "iopub.status.busy": "2025-01-29T21:54:40.296599Z",
     "iopub.status.idle": "2025-01-29T21:54:40.298933Z",
     "shell.execute_reply": "2025-01-29T21:54:40.298736Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.296742Z"
    }
   },
   "outputs": [],
   "source": [
    "errors_categorized = {}\n",
    "ERROR_TBD = \"TBD\"\n",
    "ERROR_BOTH = \"Both true and predicted labels are reasonable.\"\n",
    "ERROR_THIRD = \"Some other third label is correct.\"\n",
    "ERROR_ACTUALLY = \"The model label is correct, the true label is wrong.\"\n",
    "ERROR_SIMPLE = \"The model is just wrong.\"\n",
    "\n",
    "errors_categorized[3340] = ERROR_BOTH\n",
    "errors_categorized[3836] = ERROR_THIRD\n",
    "errors_categorized[555] = ERROR_SIMPLE\n",
    "errors_categorized[2477] = ERROR_SIMPLE\n",
    "errors_categorized[1912] = ERROR_SIMPLE\n",
    "errors_categorized[4792] = ERROR_SIMPLE\n",
    "errors_categorized[4356] = ERROR_ACTUALLY\n",
    "errors_categorized[354] = ERROR_ACTUALLY\n",
    "errors_categorized[3653] = ERROR_BOTH\n",
    "errors_categorized[3807] = ERROR_BOTH\n",
    "errors_categorized[233] = ERROR_SIMPLE\n",
    "errors_categorized[84] = ERROR_THIRD\n",
    "errors_categorized[3652] = ERROR_SIMPLE\n",
    "errors_categorized[1483] = ERROR_SIMPLE\n",
    "errors_categorized[3614] = ERROR_BOTH\n",
    "errors_categorized[3800] = ERROR_SIMPLE\n",
    "errors_categorized[5317] = ERROR_BOTH\n",
    "errors_categorized[1192] = ERROR_ACTUALLY\n",
    "errors_categorized[4313] = ERROR_SIMPLE\n",
    "errors_categorized[1172] = ERROR_ACTUALLY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "809b5f30-2e59-4353-a1e9-cd82047060a4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.299261Z",
     "iopub.status.busy": "2025-01-29T21:54:40.299202Z",
     "iopub.status.idle": "2025-01-29T21:54:40.300811Z",
     "shell.execute_reply": "2025-01-29T21:54:40.300636Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.299254Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: '0_not_relevant', 1: '1_not_happening', 2: '2_not_human', 3: '3_not_bad', 4: '4_solutions_harmful_unnecessary', 5: '5_science_unreliable', 6: '6_proponents_biased', 7: '7_fossil_fuels_needed'}\n"
     ]
    }
   ],
   "source": [
    "print(LABEL_MAPPING_INV)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "d96fbe2f-dd0e-4202-a7ad-5176aade33fd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.301161Z",
     "iopub.status.busy": "2025-01-29T21:54:40.301096Z",
     "iopub.status.idle": "2025-01-29T21:54:40.307006Z",
     "shell.execute_reply": "2025-01-29T21:54:40.306793Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.301154Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3340 Desmog None true=5_science_unreliable predict=2_not_human\n",
      "Climate is becoming increasingly warmer we hear almost every day. This is what\n",
      "has become known as Global Warming. The driving idea is that there is a linear\n",
      "relationship between CO2 increase in the atmosphere and global temperature. The\n",
      "fact, however, is that temperature has constantly gone up and down. From 1850 to\n",
      "1970, we see an almost linear relationship with Solar variability; not CO2. For\n",
      "the last 30 years, our data sets are so contaminated by personal interpretations\n",
      "and personal choices that it is almost impossible to sort up the mess in\n",
      "reliable and unreliable data,\n",
      "My categorization: Both true and predicted labels are reasonable.\n",
      "\n",
      "3836 Desmog None true=5_science_unreliable predict=2_not_human\n",
      "I think there is man made climate change, I don’t think it’s as pressing as some\n",
      "people have suggested.\n",
      "My categorization: Some other third label is correct.\n",
      "\n",
      "555 FLICC hamburg_test2 true=2_not_human predict=1_not_happening\n",
      "Global warming is real but the planet has gone through warming and cooling for\n",
      "millennia and will continue to do.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "2477 Desmog None true=4_solutions_harmful_unnecessary predict=3_not_bad\n",
      "The underlying reason blackouts are occurring is because California lacks\n",
      "reliable, in-state supply. And the reason for that is California has been\n",
      "closing both natural gas and nuclear power plants,\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "1912 Desmog None true=5_science_unreliable predict=2_not_human\n",
      "The theory of Anthropomorphic Climate Warming (AGW) rests on a three-legged\n",
      "stool. The first leg of the stool is that is that global average temperatures\n",
      "are at historic highs. The second leg of the stool is that CO2 levels are also\n",
      "at historic highs. The third leg of the stool is that peer-reviewed articles on\n",
      "global warming show a clear consensus of opinion. The recent release of e-mails\n",
      "from Britain’s Climate Research Unit, dubbed ‘climategate’ by the media, have\n",
      "made available information that kicks all three legs out from under the stool.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "4792 Desmog None true=0_not_relevant predict=3_not_bad\n",
      "That last part may be true, but it’s also true that economic development has\n",
      "made us less vulnerable, which is why there was a 99.7% decline in the death\n",
      "toll from natural disasters since its peak in 1931.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "4356 Desmog None true=4_solutions_harmful_unnecessary predict=6_proponents_biased\n",
      "I have found it necessary to go around the country pointing out that claims\n",
      "about green jobs are all phony.\n",
      "My categorization: The model label is correct, the true label is wrong.\n",
      "\n",
      "354 FLICC hamburg_test2 true=0_not_relevant predict=2_not_human\n",
      "The climate has been changing for eons!\n",
      "My categorization: The model label is correct, the true label is wrong.\n",
      "\n",
      "3653 Desmog None true=4_solutions_harmful_unnecessary predict=5_science_unreliable\n",
      "If you want to know where to stop CO2 regulation in the quest to fix the\n",
      "weather, and you’re turning the CO2 climate control knob and you want to know\n",
      "what the correct setting is, there is no answer. EPA has not supplied one. The\n",
      "court did not require them to set a standard. So there is no concentration,\n",
      "there is no global average surface temperature at which you can say ‘at last we\n",
      "have averted disaster.’ So there’s no metric, there’s no standard you can look\n",
      "to to say this much is enough… . And obviously there is no amount of regulation\n",
      "that will ever fix the weather.\n",
      "My categorization: Both true and predicted labels are reasonable.\n",
      "\n",
      "3807 Desmog None true=5_science_unreliable predict=2_not_human\n",
      "When a politician says, concerning an issue involving science, that the debate\n",
      "is over, you may be sure the debate is rolling on and not going swimmingly for\n",
      "his side. Obama is, however, quite right that climate change is a fact. The\n",
      "climate is always changing.\n",
      "My categorization: Both true and predicted labels are reasonable.\n",
      "\n",
      "233 FLICC hamburg_test3 true=1_not_happening predict=0_not_relevant\n",
      "Last time I looked the Maldives and Seychelles were still there!\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "84 FLICC CARDS true=4_solutions_harmful_unnecessary predict=5_science_unreliable\n",
      "If cloud feedback is sufficiently negative, then manmade global warming becomes\n",
      "a non-issue.\n",
      "My categorization: Some other third label is correct.\n",
      "\n",
      "3652 Desmog None true=0_not_relevant predict=6_proponents_biased\n",
      "The modern world is full of old Christian ideas gone mad.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "1483 FLICC CARDS true=4_solutions_harmful_unnecessary predict=5_science_unreliable\n",
      "The IPCC's own climate-sensitivity equations show that abating 0.06% of global\n",
      "carbon emissions would reduce CO2 concentration from a predicted business-as-\n",
      "usual 410 microatmospheres to 409.988 microatmospheres, and that this would\n",
      "reduce global mean surface temperature by just 0.00006 Celsius degrees.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "3614 Desmog None true=4_solutions_harmful_unnecessary predict=3_not_bad\n",
      "As municipalities, counties, and even countries declare a “climate emergency,”\n",
      "it is apparent that global warming is often being presented as an existential\n",
      "challenge requiring urgent and strong climate policies to avoid devastation This\n",
      "article has shown that these claims are misleading and often incorrectly\n",
      "describe the issue and its future. While climate change is real, human caused,\n",
      "and will have a mostly negative impact, it is important to remember that climate\n",
      "policies will likewise have a mostly negative impact. Thus, we must account for\n",
      "the effects of both to find the policies that will achieve the highest welfare\n",
      "gains.\n",
      "My categorization: Both true and predicted labels are reasonable.\n",
      "\n",
      "3800 Desmog None true=5_science_unreliable predict=2_not_human\n",
      "The fact of the matter is the last sixteen years doesn’t fit any model. Even\n",
      "though we continue to put a lot of CO2 in, it doesn’t fit the climate model. […]\n",
      "So something’s going on. And so I think we need to be prudent. It doesn’t mean I\n",
      "think we need to be destructive on fossil fuels.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "5317 Desmog None true=7_fossil_fuels_needed predict=4_solutions_harmful_unnecessary\n",
      "A win for Keystone XL is a defeat for the global warming movement. Green groups\n",
      "view Keystone as an opportunity to regain momentum and offset their losses after\n",
      "the death of cap-and-trade. If friends of affordable energy win this fight,\n",
      "which seems likely, the greenhouse lobby will take another hit to its prestige,\n",
      "morale, and influence eystone XL strains relations between Obama and his\n",
      "environmentalist base. If Obama approves the pipeline, greenies will be less\n",
      "motivated to work for his re-election. If he disapproves, Republicans and\n",
      "moderate Democrats will hammer him for killing job creation and increasing pain\n",
      "at the pump. Either way, the prospects for new anti-energy legislation should be\n",
      "dimmer eystone XL is bringing aging, New Lefties out of the woodwork, where they\n",
      "can misbehave and get themselves arrested.\n",
      "My categorization: Both true and predicted labels are reasonable.\n",
      "\n",
      "1192 FLICC hamburg_test1 true=0_not_relevant predict=6_proponents_biased\n",
      "She's not lecturing us, she's the mouthpiece and face of anti capitalist Marxist\n",
      "movement. She doesn't have a clue and it's not even about climate change\n",
      "My categorization: The model label is correct, the true label is wrong.\n",
      "\n",
      "4313 Desmog None true=0_not_relevant predict=1_not_happening\n",
      "Climate change will mean more extreme weather conditions and more water in the\n",
      "sea.\n",
      "My categorization: The model is just wrong.\n",
      "\n",
      "1172 FLICC Alhindi_train true=4_solutions_harmful_unnecessary predict=0_not_relevant\n",
      "“The most dramatic impacts may not be felt for 50 or 100 years.”\n",
      "My categorization: The model label is correct, the true label is wrong.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i, row in (\n",
    "    analysis_df.query(\"is_correct == 0\").sample(20, random_state=1).iterrows()\n",
    "):\n",
    "    if i not in errors_categorized:\n",
    "        errors_categorized[i] = ERROR_TBD\n",
    "    print_example(i, row, errors_categorized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "48790aa9-f053-4e90-b53d-9c6575b8b67e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.307462Z",
     "iopub.status.busy": "2025-01-29T21:54:40.307378Z",
     "iopub.status.idle": "2025-01-29T21:54:40.310509Z",
     "shell.execute_reply": "2025-01-29T21:54:40.310280Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.307455Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>error_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3340</th>\n",
       "      <td>Both true and predicted labels are reasonable.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3836</th>\n",
       "      <td>Some other third label is correct.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2477</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1912</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4792</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4356</th>\n",
       "      <td>The model label is correct, the true label is ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354</th>\n",
       "      <td>The model label is correct, the true label is ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3653</th>\n",
       "      <td>Both true and predicted labels are reasonable.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3807</th>\n",
       "      <td>Both true and predicted labels are reasonable.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>233</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>Some other third label is correct.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3652</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1483</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3614</th>\n",
       "      <td>Both true and predicted labels are reasonable.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3800</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5317</th>\n",
       "      <td>Both true and predicted labels are reasonable.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1192</th>\n",
       "      <td>The model label is correct, the true label is ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4313</th>\n",
       "      <td>The model is just wrong.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1172</th>\n",
       "      <td>The model label is correct, the true label is ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             error_type\n",
       "3340     Both true and predicted labels are reasonable.\n",
       "3836                 Some other third label is correct.\n",
       "555                            The model is just wrong.\n",
       "2477                           The model is just wrong.\n",
       "1912                           The model is just wrong.\n",
       "4792                           The model is just wrong.\n",
       "4356  The model label is correct, the true label is ...\n",
       "354   The model label is correct, the true label is ...\n",
       "3653     Both true and predicted labels are reasonable.\n",
       "3807     Both true and predicted labels are reasonable.\n",
       "233                            The model is just wrong.\n",
       "84                   Some other third label is correct.\n",
       "3652                           The model is just wrong.\n",
       "1483                           The model is just wrong.\n",
       "3614     Both true and predicted labels are reasonable.\n",
       "3800                           The model is just wrong.\n",
       "5317     Both true and predicted labels are reasonable.\n",
       "1192  The model label is correct, the true label is ...\n",
       "4313                           The model is just wrong.\n",
       "1172  The model label is correct, the true label is ..."
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "errors = pd.DataFrame(pd.Series(errors_categorized, name=\"error_type\"))\n",
    "errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "2cae7cac-f783-4b43-a9ea-30a497b6266e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.311144Z",
     "iopub.status.busy": "2025-01-29T21:54:40.310971Z",
     "iopub.status.idle": "2025-01-29T21:54:40.314613Z",
     "shell.execute_reply": "2025-01-29T21:54:40.314397Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.311134Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "error_type                                          \n",
       "The model is just wrong.                                0.45\n",
       "Both true and predicted labels are reasonable.          0.25\n",
       "The model label is correct, the true label is wrong.    0.20\n",
       "Some other third label is correct.                      0.10\n",
       "Name: count, dtype: float64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "errors.value_counts() / errors.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "a7ee528a-eea9-4a46-9270-f31bcd961add",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.315089Z",
     "iopub.status.busy": "2025-01-29T21:54:40.314999Z",
     "iopub.status.idle": "2025-01-29T21:54:40.316918Z",
     "shell.execute_reply": "2025-01-29T21:54:40.316701Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.315081Z"
    }
   },
   "outputs": [],
   "source": [
    "analysis_df[\"error_type\"] = errors[\"error_type\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "fa645016-e979-4e78-ae59-605e56870966",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.317389Z",
     "iopub.status.busy": "2025-01-29T21:54:40.317288Z",
     "iopub.status.idle": "2025-01-29T21:54:40.319658Z",
     "shell.execute_reply": "2025-01-29T21:54:40.319380Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.317381Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "error_type\n",
       "NaN                                                     1199\n",
       "The model is just wrong.                                   9\n",
       "Both true and predicted labels are reasonable.             5\n",
       "The model label is correct, the true label is wrong.       4\n",
       "Some other third label is correct.                         2\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df[\"error_type\"].value_counts(dropna=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "116b183b-0af3-4055-ad52-a2c624c231be",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-29T21:54:40.320094Z",
     "iopub.status.busy": "2025-01-29T21:54:40.319990Z",
     "iopub.status.idle": "2025-01-29T21:54:40.324988Z",
     "shell.execute_reply": "2025-01-29T21:54:40.324788Z",
     "shell.execute_reply.started": "2025-01-29T21:54:40.320086Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>quote</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <th>error_type</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">Desmog</th>\n",
       "      <th>Both true and predicted labels are reasonable.</th>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Some other third label is correct.</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The model is just wrong.</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The model label is correct, the true label is wrong.</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NaN</th>\n",
       "      <td>109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">FLICC</th>\n",
       "      <th>Some other third label is correct.</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The model is just wrong.</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The model label is correct, the true label is wrong.</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NaN</th>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                           quote\n",
       "source error_type                                               \n",
       "Desmog Both true and predicted labels are reasonable.          5\n",
       "       Some other third label is correct.                      1\n",
       "       The model is just wrong.                                6\n",
       "       The model label is correct, the true label is w...      1\n",
       "       NaN                                                   109\n",
       "FLICC  Some other third label is correct.                      1\n",
       "       The model is just wrong.                                3\n",
       "       The model label is correct, the true label is w...      3\n",
       "       NaN                                                    44"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "analysis_df.query(\"is_correct == 0\").groupby(\n",
    "    [\"source\", \"error_type\"], dropna=False\n",
    ").agg({\"quote\": \"count\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f728694c-dc2e-472e-b148-693ecbf220a7",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.13.1"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}