Hugging Face's logo

Datasets:
imageomics
/
IDLE-OO-Camera-Traps

Tasks:
Image Classification
Zero-Shot Classification
Modalities:
Image
Text
Formats:
imagefolder
Languages:
English
Latin
Size:
1K - 10K
ArXiv:
Tags:
biology
image
imageomics
animals
CV
balanced
DOI:
Libraries:
Datasets
Dataset card Data Studio Files
xet
Community
9
egrace479 commited on
Commit
76411b3
·
verified ·
1 Parent(s): 99fef5e

Add notebooks generating the CSVs in data/potential-test-sets/filtered

Browse files
Files changed (4) hide show
  1. notebooks/lilabc_test-ENA24.ipynb +0 -0
  2. notebooks/lilabc_test-desert-lion.ipynb +0 -0
  3. notebooks/lilabc_test-ohio-small-animal.ipynb +1662 -0
  4. notebooks/lilabc_test-orinoquia.ipynb +0 -0
notebooks/lilabc_test-ENA24.ipynb ADDED
The diff for this file is too large to render. See raw diff
 
notebooks/lilabc_test-desert-lion.ipynb ADDED
The diff for this file is too large to render. See raw diff
 
notebooks/lilabc_test-ohio-small-animal.ipynb ADDED
@@ -0,0 +1,1662 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cells": [
3
+ {
4
+ "cell_type": "code",
5
+ "execution_count": 1,
6
+ "metadata": {},
7
+ "outputs": [],
8
+ "source": [
9
+ "import pandas as pd\n",
10
+ "import seaborn as sns\n",
11
+ "\n",
12
+ "sns.set_style(\"whitegrid\")"
13
+ ]
14
+ },
15
+ {
16
+ "cell_type": "markdown",
17
+ "metadata": {},
18
+ "source": [
19
+ "Reading in our potential test sets from [here](https://huggingface.co/datasets/imageomics/lila-bc-camera/tree/d18307b285217d18b31d1a7b2c9091bb0873ade0/data/potential-test-sets):\n",
20
+ " - [Ohio Small Animals](https://lila.science/datasets/ohio-small-animals/)\n",
21
+ " - [Desert Lion Conservation Camera Traps](https://lila.science/datasets/desert-lion-conservation-camera-traps/)\n",
22
+ " - [Orinoquia Camera Traps](https://lila.science/datasets/orinoquia-camera-traps/)\n",
23
+ " - [Island Conservation Camera Traps](https://lila.science/datasets/island-conservation-camera-traps/)\n",
24
+ " - [ENA24](https://lila.science/datasets/ena24detection)\n",
25
+ "\n",
26
+ "We'll clean them down to just the taxa and identifier columns, then further reduce to make balanced test sets for each.\n",
27
+ "\n",
28
+ "# Ohio Small Animals Datasets\n",
29
+ "upper/lower bounded and balanced"
30
+ ]
31
+ },
32
+ {
33
+ "cell_type": "code",
34
+ "execution_count": 2,
35
+ "metadata": {},
36
+ "outputs": [
37
+ {
38
+ "data": {
39
+ "text/html": [
40
+ "<div>\n",
41
+ "<style scoped>\n",
42
+ " .dataframe tbody tr th:only-of-type {\n",
43
+ " vertical-align: middle;\n",
44
+ " }\n",
45
+ "\n",
46
+ " .dataframe tbody tr th {\n",
47
+ " vertical-align: top;\n",
48
+ " }\n",
49
+ "\n",
50
+ " .dataframe thead th {\n",
51
+ " text-align: right;\n",
52
+ " }\n",
53
+ "</style>\n",
54
+ "<table border=\"1\" class=\"dataframe\">\n",
55
+ " <thead>\n",
56
+ " <tr style=\"text-align: right;\">\n",
57
+ " <th></th>\n",
58
+ " <th>dataset_name</th>\n",
59
+ " <th>url_gcp</th>\n",
60
+ " <th>url_aws</th>\n",
61
+ " <th>url_azure</th>\n",
62
+ " <th>image_id</th>\n",
63
+ " <th>sequence_id</th>\n",
64
+ " <th>location_id</th>\n",
65
+ " <th>frame_num</th>\n",
66
+ " <th>original_label</th>\n",
67
+ " <th>scientific_name</th>\n",
68
+ " <th>...</th>\n",
69
+ " <th>superfamily</th>\n",
70
+ " <th>family</th>\n",
71
+ " <th>subfamily</th>\n",
72
+ " <th>tribe</th>\n",
73
+ " <th>genus</th>\n",
74
+ " <th>species</th>\n",
75
+ " <th>subspecies</th>\n",
76
+ " <th>variety</th>\n",
77
+ " <th>multi_species</th>\n",
78
+ " <th>num_species</th>\n",
79
+ " </tr>\n",
80
+ " </thead>\n",
81
+ " <tbody>\n",
82
+ " <tr>\n",
83
+ " <th>0</th>\n",
84
+ " <td>Ohio Small Animals</td>\n",
85
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
86
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
87
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
88
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
89
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
90
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
91
+ " <td>0</td>\n",
92
+ " <td>american_bullfrog</td>\n",
93
+ " <td>lithobates catesbeianus</td>\n",
94
+ " <td>...</td>\n",
95
+ " <td>NaN</td>\n",
96
+ " <td>ranidae</td>\n",
97
+ " <td>NaN</td>\n",
98
+ " <td>NaN</td>\n",
99
+ " <td>lithobates</td>\n",
100
+ " <td>lithobates catesbeianus</td>\n",
101
+ " <td>NaN</td>\n",
102
+ " <td>NaN</td>\n",
103
+ " <td>False</td>\n",
104
+ " <td>1.0</td>\n",
105
+ " </tr>\n",
106
+ " <tr>\n",
107
+ " <th>1</th>\n",
108
+ " <td>Ohio Small Animals</td>\n",
109
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
110
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
111
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
112
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
113
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
114
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
115
+ " <td>1</td>\n",
116
+ " <td>american_bullfrog</td>\n",
117
+ " <td>lithobates catesbeianus</td>\n",
118
+ " <td>...</td>\n",
119
+ " <td>NaN</td>\n",
120
+ " <td>ranidae</td>\n",
121
+ " <td>NaN</td>\n",
122
+ " <td>NaN</td>\n",
123
+ " <td>lithobates</td>\n",
124
+ " <td>lithobates catesbeianus</td>\n",
125
+ " <td>NaN</td>\n",
126
+ " <td>NaN</td>\n",
127
+ " <td>False</td>\n",
128
+ " <td>1.0</td>\n",
129
+ " </tr>\n",
130
+ " <tr>\n",
131
+ " <th>2</th>\n",
132
+ " <td>Ohio Small Animals</td>\n",
133
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
134
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
135
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
136
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
137
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
138
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
139
+ " <td>2</td>\n",
140
+ " <td>american_bullfrog</td>\n",
141
+ " <td>lithobates catesbeianus</td>\n",
142
+ " <td>...</td>\n",
143
+ " <td>NaN</td>\n",
144
+ " <td>ranidae</td>\n",
145
+ " <td>NaN</td>\n",
146
+ " <td>NaN</td>\n",
147
+ " <td>lithobates</td>\n",
148
+ " <td>lithobates catesbeianus</td>\n",
149
+ " <td>NaN</td>\n",
150
+ " <td>NaN</td>\n",
151
+ " <td>False</td>\n",
152
+ " <td>1.0</td>\n",
153
+ " </tr>\n",
154
+ " <tr>\n",
155
+ " <th>3</th>\n",
156
+ " <td>Ohio Small Animals</td>\n",
157
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
158
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
159
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
160
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
161
+ " <td>Ohio Small Animals : location_BIWA4N_sequence_...</td>\n",
162
+ " <td>Ohio Small Animals : BIWA4N</td>\n",
163
+ " <td>0</td>\n",
164
+ " <td>american_bullfrog</td>\n",
165
+ " <td>lithobates catesbeianus</td>\n",
166
+ " <td>...</td>\n",
167
+ " <td>NaN</td>\n",
168
+ " <td>ranidae</td>\n",
169
+ " <td>NaN</td>\n",
170
+ " <td>NaN</td>\n",
171
+ " <td>lithobates</td>\n",
172
+ " <td>lithobates catesbeianus</td>\n",
173
+ " <td>NaN</td>\n",
174
+ " <td>NaN</td>\n",
175
+ " <td>False</td>\n",
176
+ " <td>1.0</td>\n",
177
+ " </tr>\n",
178
+ " <tr>\n",
179
+ " <th>4</th>\n",
180
+ " <td>Ohio Small Animals</td>\n",
181
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
182
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
183
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
184
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
185
+ " <td>Ohio Small Animals : location_BIWA4N_sequence_...</td>\n",
186
+ " <td>Ohio Small Animals : BIWA4N</td>\n",
187
+ " <td>1</td>\n",
188
+ " <td>american_bullfrog</td>\n",
189
+ " <td>lithobates catesbeianus</td>\n",
190
+ " <td>...</td>\n",
191
+ " <td>NaN</td>\n",
192
+ " <td>ranidae</td>\n",
193
+ " <td>NaN</td>\n",
194
+ " <td>NaN</td>\n",
195
+ " <td>lithobates</td>\n",
196
+ " <td>lithobates catesbeianus</td>\n",
197
+ " <td>NaN</td>\n",
198
+ " <td>NaN</td>\n",
199
+ " <td>False</td>\n",
200
+ " <td>1.0</td>\n",
201
+ " </tr>\n",
202
+ " </tbody>\n",
203
+ "</table>\n",
204
+ "<p>5 rows × 34 columns</p>\n",
205
+ "</div>"
206
+ ],
207
+ "text/plain": [
208
+ " dataset_name url_gcp \\\n",
209
+ "0 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
210
+ "1 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
211
+ "2 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
212
+ "3 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
213
+ "4 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
214
+ "\n",
215
+ " url_aws \\\n",
216
+ "0 http://us-west-2.opendata.source.coop.s3.amazo... \n",
217
+ "1 http://us-west-2.opendata.source.coop.s3.amazo... \n",
218
+ "2 http://us-west-2.opendata.source.coop.s3.amazo... \n",
219
+ "3 http://us-west-2.opendata.source.coop.s3.amazo... \n",
220
+ "4 http://us-west-2.opendata.source.coop.s3.amazo... \n",
221
+ "\n",
222
+ " url_azure \\\n",
223
+ "0 https://lilawildlife.blob.core.windows.net/lil... \n",
224
+ "1 https://lilawildlife.blob.core.windows.net/lil... \n",
225
+ "2 https://lilawildlife.blob.core.windows.net/lil... \n",
226
+ "3 https://lilawildlife.blob.core.windows.net/lil... \n",
227
+ "4 https://lilawildlife.blob.core.windows.net/lil... \n",
228
+ "\n",
229
+ " image_id \\\n",
230
+ "0 Ohio Small Animals : Images/Sorted_by_species/... \n",
231
+ "1 Ohio Small Animals : Images/Sorted_by_species/... \n",
232
+ "2 Ohio Small Animals : Images/Sorted_by_species/... \n",
233
+ "3 Ohio Small Animals : Images/Sorted_by_species/... \n",
234
+ "4 Ohio Small Animals : Images/Sorted_by_species/... \n",
235
+ "\n",
236
+ " sequence_id \\\n",
237
+ "0 Ohio Small Animals : location_BIWA3N_sequence_... \n",
238
+ "1 Ohio Small Animals : location_BIWA3N_sequence_... \n",
239
+ "2 Ohio Small Animals : location_BIWA3N_sequence_... \n",
240
+ "3 Ohio Small Animals : location_BIWA4N_sequence_... \n",
241
+ "4 Ohio Small Animals : location_BIWA4N_sequence_... \n",
242
+ "\n",
243
+ " location_id frame_num original_label \\\n",
244
+ "0 Ohio Small Animals : BIWA3N 0 american_bullfrog \n",
245
+ "1 Ohio Small Animals : BIWA3N 1 american_bullfrog \n",
246
+ "2 Ohio Small Animals : BIWA3N 2 american_bullfrog \n",
247
+ "3 Ohio Small Animals : BIWA4N 0 american_bullfrog \n",
248
+ "4 Ohio Small Animals : BIWA4N 1 american_bullfrog \n",
249
+ "\n",
250
+ " scientific_name ... superfamily family subfamily tribe \\\n",
251
+ "0 lithobates catesbeianus ... NaN ranidae NaN NaN \n",
252
+ "1 lithobates catesbeianus ... NaN ranidae NaN NaN \n",
253
+ "2 lithobates catesbeianus ... NaN ranidae NaN NaN \n",
254
+ "3 lithobates catesbeianus ... NaN ranidae NaN NaN \n",
255
+ "4 lithobates catesbeianus ... NaN ranidae NaN NaN \n",
256
+ "\n",
257
+ " genus species subspecies variety multi_species \\\n",
258
+ "0 lithobates lithobates catesbeianus NaN NaN False \n",
259
+ "1 lithobates lithobates catesbeianus NaN NaN False \n",
260
+ "2 lithobates lithobates catesbeianus NaN NaN False \n",
261
+ "3 lithobates lithobates catesbeianus NaN NaN False \n",
262
+ "4 lithobates lithobates catesbeianus NaN NaN False \n",
263
+ "\n",
264
+ " num_species \n",
265
+ "0 1.0 \n",
266
+ "1 1.0 \n",
267
+ "2 1.0 \n",
268
+ "3 1.0 \n",
269
+ "4 1.0 \n",
270
+ "\n",
271
+ "[5 rows x 34 columns]"
272
+ ]
273
+ },
274
+ "execution_count": 2,
275
+ "metadata": {},
276
+ "output_type": "execute_result"
277
+ }
278
+ ],
279
+ "source": [
280
+ "df = pd.read_csv(\"../data/potential-test-sets/Ohio_Small_Animals_image_urls_and_labels.csv\", low_memory = False)\n",
281
+ "df.head()"
282
+ ]
283
+ },
284
+ {
285
+ "cell_type": "code",
286
+ "execution_count": 3,
287
+ "metadata": {},
288
+ "outputs": [
289
+ {
290
+ "data": {
291
+ "text/plain": [
292
+ "Index(['dataset_name', 'url_gcp', 'url_aws', 'url_azure', 'image_id',\n",
293
+ " 'sequence_id', 'location_id', 'frame_num', 'original_label',\n",
294
+ " 'scientific_name', 'common_name', 'datetime', 'annotation_level',\n",
295
+ " 'kingdom', 'phylum', 'subphylum', 'superclass', 'class', 'subclass',\n",
296
+ " 'infraclass', 'superorder', 'order', 'suborder', 'infraorder',\n",
297
+ " 'superfamily', 'family', 'subfamily', 'tribe', 'genus', 'species',\n",
298
+ " 'subspecies', 'variety', 'multi_species', 'num_species'],\n",
299
+ " dtype='object')"
300
+ ]
301
+ },
302
+ "execution_count": 3,
303
+ "metadata": {},
304
+ "output_type": "execute_result"
305
+ }
306
+ ],
307
+ "source": [
308
+ "df.columns"
309
+ ]
310
+ },
311
+ {
312
+ "cell_type": "markdown",
313
+ "metadata": {},
314
+ "source": [
315
+ "Observe that we also now get multiple URL options; `url_aws` will likely be best/fastest for use with [`distributed-downloader`](https://github.com/Imageomics/distributed-downloader) to get the images."
316
+ ]
317
+ },
318
+ {
319
+ "cell_type": "code",
320
+ "execution_count": 4,
321
+ "metadata": {},
322
+ "outputs": [
323
+ {
324
+ "name": "stdout",
325
+ "output_type": "stream",
326
+ "text": [
327
+ "<class 'pandas.core.frame.DataFrame'>\n",
328
+ "RangeIndex: 107106 entries, 0 to 107105\n",
329
+ "Data columns (total 34 columns):\n",
330
+ " # Column Non-Null Count Dtype \n",
331
+ "--- ------ -------------- ----- \n",
332
+ " 0 dataset_name 107106 non-null object \n",
333
+ " 1 url_gcp 107106 non-null object \n",
334
+ " 2 url_aws 107106 non-null object \n",
335
+ " 3 url_azure 107106 non-null object \n",
336
+ " 4 image_id 107106 non-null object \n",
337
+ " 5 sequence_id 107106 non-null object \n",
338
+ " 6 location_id 107106 non-null object \n",
339
+ " 7 frame_num 107106 non-null int64 \n",
340
+ " 8 original_label 107106 non-null object \n",
341
+ " 9 scientific_name 107106 non-null object \n",
342
+ " 10 common_name 107106 non-null object \n",
343
+ " 11 datetime 107106 non-null object \n",
344
+ " 12 annotation_level 107106 non-null object \n",
345
+ " 13 kingdom 107106 non-null object \n",
346
+ " 14 phylum 102031 non-null object \n",
347
+ " 15 subphylum 102031 non-null object \n",
348
+ " 16 superclass 0 non-null float64\n",
349
+ " 17 class 102031 non-null object \n",
350
+ " 18 subclass 38900 non-null object \n",
351
+ " 19 infraclass 38900 non-null object \n",
352
+ " 20 superorder 37915 non-null object \n",
353
+ " 21 order 102031 non-null object \n",
354
+ " 22 suborder 67797 non-null object \n",
355
+ " 23 infraorder 0 non-null float64\n",
356
+ " 24 superfamily 26395 non-null object \n",
357
+ " 25 family 102031 non-null object \n",
358
+ " 26 subfamily 75088 non-null object \n",
359
+ " 27 tribe 31051 non-null object \n",
360
+ " 28 genus 102031 non-null object \n",
361
+ " 29 species 102031 non-null object \n",
362
+ " 30 subspecies 38087 non-null object \n",
363
+ " 31 variety 0 non-null float64\n",
364
+ " 32 multi_species 107106 non-null bool \n",
365
+ " 33 num_species 107106 non-null float64\n",
366
+ "dtypes: bool(1), float64(4), int64(1), object(28)\n",
367
+ "memory usage: 27.1+ MB\n"
368
+ ]
369
+ }
370
+ ],
371
+ "source": [
372
+ "df.info(show_counts = True)"
373
+ ]
374
+ },
375
+ {
376
+ "cell_type": "markdown",
377
+ "metadata": {},
378
+ "source": [
379
+ "This is mostly filled in. Let's get some counts on the taxa, then start looking at balancing it."
380
+ ]
381
+ },
382
+ {
383
+ "cell_type": "code",
384
+ "execution_count": 7,
385
+ "metadata": {},
386
+ "outputs": [],
387
+ "source": [
388
+ "lin_taxa = ['kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']\n",
389
+ "taxa_cols = ['original_label', 'scientific_name', 'common_name', 'kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']"
390
+ ]
391
+ },
392
+ {
393
+ "cell_type": "code",
394
+ "execution_count": 8,
395
+ "metadata": {},
396
+ "outputs": [
397
+ {
398
+ "data": {
399
+ "text/plain": [
400
+ "original_label 45\n",
401
+ "scientific_name 45\n",
402
+ "common_name 45\n",
403
+ "kingdom 1\n",
404
+ "phylum 1\n",
405
+ "class 4\n",
406
+ "order 10\n",
407
+ "family 25\n",
408
+ "genus 37\n",
409
+ "species 44\n",
410
+ "dtype: int64"
411
+ ]
412
+ },
413
+ "execution_count": 8,
414
+ "metadata": {},
415
+ "output_type": "execute_result"
416
+ }
417
+ ],
418
+ "source": [
419
+ "df[taxa_cols].nunique()"
420
+ ]
421
+ },
422
+ {
423
+ "cell_type": "markdown",
424
+ "metadata": {},
425
+ "source": [
426
+ "We only have 45--there's also null species values, so that'd be the distinction there.\n",
427
+ "\n",
428
+ "Also, should check for duplicated images."
429
+ ]
430
+ },
431
+ {
432
+ "cell_type": "code",
433
+ "execution_count": 9,
434
+ "metadata": {},
435
+ "outputs": [
436
+ {
437
+ "name": "stdout",
438
+ "output_type": "stream",
439
+ "text": [
440
+ "number of unique images: 107106\n"
441
+ ]
442
+ },
443
+ {
444
+ "data": {
445
+ "text/plain": [
446
+ "multi_species\n",
447
+ "False 107106\n",
448
+ "Name: count, dtype: int64"
449
+ ]
450
+ },
451
+ "execution_count": 9,
452
+ "metadata": {},
453
+ "output_type": "execute_result"
454
+ }
455
+ ],
456
+ "source": [
457
+ "print(f\"number of unique images: {df[\"image_id\"].nunique()}\")\n",
458
+ "\n",
459
+ "df[\"multi_species\"].value_counts()"
460
+ ]
461
+ },
462
+ {
463
+ "cell_type": "markdown",
464
+ "metadata": {},
465
+ "source": [
466
+ "Okay, none of these have multiple species per image, so now let's look at how many images we have per species."
467
+ ]
468
+ },
469
+ {
470
+ "cell_type": "code",
471
+ "execution_count": 10,
472
+ "metadata": {},
473
+ "outputs": [
474
+ {
475
+ "data": {
476
+ "text/html": [
477
+ "<div>\n",
478
+ "<style scoped>\n",
479
+ " .dataframe tbody tr th:only-of-type {\n",
480
+ " vertical-align: middle;\n",
481
+ " }\n",
482
+ "\n",
483
+ " .dataframe tbody tr th {\n",
484
+ " vertical-align: top;\n",
485
+ " }\n",
486
+ "\n",
487
+ " .dataframe thead th {\n",
488
+ " text-align: right;\n",
489
+ " }\n",
490
+ "</style>\n",
491
+ "<table border=\"1\" class=\"dataframe\">\n",
492
+ " <thead>\n",
493
+ " <tr style=\"text-align: right;\">\n",
494
+ " <th></th>\n",
495
+ " <th>dataset_name</th>\n",
496
+ " <th>url_gcp</th>\n",
497
+ " <th>url_aws</th>\n",
498
+ " <th>url_azure</th>\n",
499
+ " <th>image_id</th>\n",
500
+ " <th>sequence_id</th>\n",
501
+ " <th>location_id</th>\n",
502
+ " <th>frame_num</th>\n",
503
+ " <th>original_label</th>\n",
504
+ " <th>scientific_name</th>\n",
505
+ " <th>...</th>\n",
506
+ " <th>superfamily</th>\n",
507
+ " <th>family</th>\n",
508
+ " <th>subfamily</th>\n",
509
+ " <th>tribe</th>\n",
510
+ " <th>genus</th>\n",
511
+ " <th>species</th>\n",
512
+ " <th>subspecies</th>\n",
513
+ " <th>variety</th>\n",
514
+ " <th>multi_species</th>\n",
515
+ " <th>num_species</th>\n",
516
+ " </tr>\n",
517
+ " </thead>\n",
518
+ " <tbody>\n",
519
+ " <tr>\n",
520
+ " <th>23682</th>\n",
521
+ " <td>Ohio Small Animals</td>\n",
522
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
523
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
524
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
525
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
526
+ " <td>Ohio Small Animals : location_GRN1_sequence_in...</td>\n",
527
+ " <td>Ohio Small Animals : GRN1</td>\n",
528
+ " <td>1</td>\n",
529
+ " <td>invertebrate</td>\n",
530
+ " <td>animalia</td>\n",
531
+ " <td>...</td>\n",
532
+ " <td>NaN</td>\n",
533
+ " <td>NaN</td>\n",
534
+ " <td>NaN</td>\n",
535
+ " <td>NaN</td>\n",
536
+ " <td>NaN</td>\n",
537
+ " <td>NaN</td>\n",
538
+ " <td>NaN</td>\n",
539
+ " <td>NaN</td>\n",
540
+ " <td>False</td>\n",
541
+ " <td>1.0</td>\n",
542
+ " </tr>\n",
543
+ " <tr>\n",
544
+ " <th>24727</th>\n",
545
+ " <td>Ohio Small Animals</td>\n",
546
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
547
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
548
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
549
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
550
+ " <td>Ohio Small Animals : location_MLN3_sequence_in...</td>\n",
551
+ " <td>Ohio Small Animals : MLN3</td>\n",
552
+ " <td>1</td>\n",
553
+ " <td>invertebrate</td>\n",
554
+ " <td>animalia</td>\n",
555
+ " <td>...</td>\n",
556
+ " <td>NaN</td>\n",
557
+ " <td>NaN</td>\n",
558
+ " <td>NaN</td>\n",
559
+ " <td>NaN</td>\n",
560
+ " <td>NaN</td>\n",
561
+ " <td>NaN</td>\n",
562
+ " <td>NaN</td>\n",
563
+ " <td>NaN</td>\n",
564
+ " <td>False</td>\n",
565
+ " <td>1.0</td>\n",
566
+ " </tr>\n",
567
+ " <tr>\n",
568
+ " <th>26630</th>\n",
569
+ " <td>Ohio Small Animals</td>\n",
570
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
571
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
572
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
573
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
574
+ " <td>Ohio Small Animals : location_WIL3_sequence_in...</td>\n",
575
+ " <td>Ohio Small Animals : WIL3</td>\n",
576
+ " <td>0</td>\n",
577
+ " <td>invertebrate</td>\n",
578
+ " <td>animalia</td>\n",
579
+ " <td>...</td>\n",
580
+ " <td>NaN</td>\n",
581
+ " <td>NaN</td>\n",
582
+ " <td>NaN</td>\n",
583
+ " <td>NaN</td>\n",
584
+ " <td>NaN</td>\n",
585
+ " <td>NaN</td>\n",
586
+ " <td>NaN</td>\n",
587
+ " <td>NaN</td>\n",
588
+ " <td>False</td>\n",
589
+ " <td>1.0</td>\n",
590
+ " </tr>\n",
591
+ " <tr>\n",
592
+ " <th>26769</th>\n",
593
+ " <td>Ohio Small Animals</td>\n",
594
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
595
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
596
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
597
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
598
+ " <td>Ohio Small Animals : location_YWN1_sequence_in...</td>\n",
599
+ " <td>Ohio Small Animals : YWN1</td>\n",
600
+ " <td>2</td>\n",
601
+ " <td>invertebrate</td>\n",
602
+ " <td>animalia</td>\n",
603
+ " <td>...</td>\n",
604
+ " <td>NaN</td>\n",
605
+ " <td>NaN</td>\n",
606
+ " <td>NaN</td>\n",
607
+ " <td>NaN</td>\n",
608
+ " <td>NaN</td>\n",
609
+ " <td>NaN</td>\n",
610
+ " <td>NaN</td>\n",
611
+ " <td>NaN</td>\n",
612
+ " <td>False</td>\n",
613
+ " <td>1.0</td>\n",
614
+ " </tr>\n",
615
+ " </tbody>\n",
616
+ "</table>\n",
617
+ "<p>4 rows × 34 columns</p>\n",
618
+ "</div>"
619
+ ],
620
+ "text/plain": [
621
+ " dataset_name url_gcp \\\n",
622
+ "23682 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
623
+ "24727 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
624
+ "26630 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
625
+ "26769 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
626
+ "\n",
627
+ " url_aws \\\n",
628
+ "23682 http://us-west-2.opendata.source.coop.s3.amazo... \n",
629
+ "24727 http://us-west-2.opendata.source.coop.s3.amazo... \n",
630
+ "26630 http://us-west-2.opendata.source.coop.s3.amazo... \n",
631
+ "26769 http://us-west-2.opendata.source.coop.s3.amazo... \n",
632
+ "\n",
633
+ " url_azure \\\n",
634
+ "23682 https://lilawildlife.blob.core.windows.net/lil... \n",
635
+ "24727 https://lilawildlife.blob.core.windows.net/lil... \n",
636
+ "26630 https://lilawildlife.blob.core.windows.net/lil... \n",
637
+ "26769 https://lilawildlife.blob.core.windows.net/lil... \n",
638
+ "\n",
639
+ " image_id \\\n",
640
+ "23682 Ohio Small Animals : Images/Sorted_by_species/... \n",
641
+ "24727 Ohio Small Animals : Images/Sorted_by_species/... \n",
642
+ "26630 Ohio Small Animals : Images/Sorted_by_species/... \n",
643
+ "26769 Ohio Small Animals : Images/Sorted_by_species/... \n",
644
+ "\n",
645
+ " sequence_id \\\n",
646
+ "23682 Ohio Small Animals : location_GRN1_sequence_in... \n",
647
+ "24727 Ohio Small Animals : location_MLN3_sequence_in... \n",
648
+ "26630 Ohio Small Animals : location_WIL3_sequence_in... \n",
649
+ "26769 Ohio Small Animals : location_YWN1_sequence_in... \n",
650
+ "\n",
651
+ " location_id frame_num original_label scientific_name \\\n",
652
+ "23682 Ohio Small Animals : GRN1 1 invertebrate animalia \n",
653
+ "24727 Ohio Small Animals : MLN3 1 invertebrate animalia \n",
654
+ "26630 Ohio Small Animals : WIL3 0 invertebrate animalia \n",
655
+ "26769 Ohio Small Animals : YWN1 2 invertebrate animalia \n",
656
+ "\n",
657
+ " ... superfamily family subfamily tribe genus species subspecies \\\n",
658
+ "23682 ... NaN NaN NaN NaN NaN NaN NaN \n",
659
+ "24727 ... NaN NaN NaN NaN NaN NaN NaN \n",
660
+ "26630 ... NaN NaN NaN NaN NaN NaN NaN \n",
661
+ "26769 ... NaN NaN NaN NaN NaN NaN NaN \n",
662
+ "\n",
663
+ " variety multi_species num_species \n",
664
+ "23682 NaN False 1.0 \n",
665
+ "24727 NaN False 1.0 \n",
666
+ "26630 NaN False 1.0 \n",
667
+ "26769 NaN False 1.0 \n",
668
+ "\n",
669
+ "[4 rows x 34 columns]"
670
+ ]
671
+ },
672
+ "execution_count": 10,
673
+ "metadata": {},
674
+ "output_type": "execute_result"
675
+ }
676
+ ],
677
+ "source": [
678
+ "df.loc[df[\"species\"].isna()].sample(4)"
679
+ ]
680
+ },
681
+ {
682
+ "cell_type": "markdown",
683
+ "metadata": {},
684
+ "source": [
685
+ "Ah, these are probably not images we want."
686
+ ]
687
+ },
688
+ {
689
+ "cell_type": "code",
690
+ "execution_count": 11,
691
+ "metadata": {},
692
+ "outputs": [
693
+ {
694
+ "data": {
695
+ "text/plain": [
696
+ "scientific_name\n",
697
+ "thamnophis sirtalis sirtalis 31899\n",
698
+ "melospiza melodia 14567\n",
699
+ "microtus pennsylvanicus 14169\n",
700
+ "peromyscus leucopus 10548\n",
701
+ "troglodytes aedon aedon 5934\n",
702
+ "animalia 5075\n",
703
+ "plestiodon fasciatus 5045\n",
704
+ "sorex cinereus 4242\n",
705
+ "sylvilagus floridanus 3263\n",
706
+ "neogale frenata 2325\n",
707
+ "napaeozapus insignis 1510\n",
708
+ "thamnophis radix 1272\n",
709
+ "sistrurus catenatus 1189\n",
710
+ "didelphis virginiana 985\n",
711
+ "geothlypis trichas 802\n",
712
+ "Name: count, dtype: int64"
713
+ ]
714
+ },
715
+ "execution_count": 11,
716
+ "metadata": {},
717
+ "output_type": "execute_result"
718
+ }
719
+ ],
720
+ "source": [
721
+ "df.scientific_name.value_counts()[:15]"
722
+ ]
723
+ },
724
+ {
725
+ "cell_type": "code",
726
+ "execution_count": 13,
727
+ "metadata": {},
728
+ "outputs": [
729
+ {
730
+ "data": {
731
+ "text/plain": [
732
+ "scientific_name\n",
733
+ "pantherophis spiloides 68\n",
734
+ "heterodon platirhinos 67\n",
735
+ "procyon lotor 62\n",
736
+ "lithobates clamitans 47\n",
737
+ "marmota monax 44\n",
738
+ "clonophis kirtlandii 44\n",
739
+ "passerina cyanea 23\n",
740
+ "chrysemys picta 23\n",
741
+ "porzana carolina 13\n",
742
+ "lithobates catesbeianus 12\n",
743
+ "dumetella carolinensis 10\n",
744
+ "storeria occipitomaculata 9\n",
745
+ "rattus norvegicus 8\n",
746
+ "chelydra serpentina 6\n",
747
+ "sialia sialis 1\n",
748
+ "Name: count, dtype: int64"
749
+ ]
750
+ },
751
+ "execution_count": 13,
752
+ "metadata": {},
753
+ "output_type": "execute_result"
754
+ }
755
+ ],
756
+ "source": [
757
+ "df.scientific_name.value_counts()[30:]"
758
+ ]
759
+ },
760
+ {
761
+ "cell_type": "markdown",
762
+ "metadata": {},
763
+ "source": [
764
+ "Looking at least-represented species:\n",
765
+ " - _lithobates catesbeianus_ isn't endangered itself, but is problematic for native species [USFWS Assessment](https://www.fws.gov/sites/default/files/documents/Ecological-Risk-Screening-Summary-American-Bullfrog.pdf), so worth testing on.\n",
766
+ " - _dumetella carolinensis_ listed as stable, least concern by [IUCN Redlist](https://www.iucnredlist.org/species/22711013/94272855), okay to exclude for balance.\n",
767
+ " - Similarly for _rattus norvegicus_ [IUCN](https://www.iucnredlist.org/species/19353/165118026), _chelydra serpentina_ is considered decreasing (still LC though) [IUCN](https://www.iucnredlist.org/species/163424/251347989), _sialia sialis_ increasing [IUCN](https://www.iucnredlist.org/species/22708550/139388955).\n",
768
+ " - _storeria occipitomaculata_ stable of least concern, though it \"needs updating\" [IUCN](https://www.iucnredlist.org/species/63930/12729296).\n",
769
+ "\n",
770
+ "So, for the balanced set we'll remove those that aren't classified beyond kingdom (all of which do have down to species), then randomly select 12 per species (to catch _lithobates catesbeianus_). That'd give us a 468 image test set (39 species at 12 images per species)."
771
+ ]
772
+ },
773
+ {
774
+ "cell_type": "code",
775
+ "execution_count": 14,
776
+ "metadata": {},
777
+ "outputs": [
778
+ {
779
+ "data": {
780
+ "text/plain": [
781
+ "<Axes: xlabel='Count', ylabel='class'>"
782
+ ]
783
+ },
784
+ "execution_count": 14,
785
+ "metadata": {},
786
+ "output_type": "execute_result"
787
+ },
788
+ {
789
+ "data": {
790
+ "image/png": "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",
791
+ "text/plain": [
792
+ "<Figure size 640x480 with 1 Axes>"
793
+ ]
794
+ },
795
+ "metadata": {},
796
+ "output_type": "display_data"
797
+ }
798
+ ],
799
+ "source": [
800
+ "sns.histplot(df, y = 'class')"
801
+ ]
802
+ },
803
+ {
804
+ "cell_type": "markdown",
805
+ "metadata": {},
806
+ "source": [
807
+ "### Filter Kingdom-only classifications"
808
+ ]
809
+ },
810
+ {
811
+ "cell_type": "code",
812
+ "execution_count": 15,
813
+ "metadata": {},
814
+ "outputs": [
815
+ {
816
+ "name": "stdout",
817
+ "output_type": "stream",
818
+ "text": [
819
+ "<class 'pandas.core.frame.DataFrame'>\n",
820
+ "Index: 102031 entries, 0 to 107105\n",
821
+ "Data columns (total 10 columns):\n",
822
+ " # Column Non-Null Count Dtype \n",
823
+ "--- ------ -------------- ----- \n",
824
+ " 0 original_label 102031 non-null object\n",
825
+ " 1 scientific_name 102031 non-null object\n",
826
+ " 2 common_name 102031 non-null object\n",
827
+ " 3 kingdom 102031 non-null object\n",
828
+ " 4 phylum 102031 non-null object\n",
829
+ " 5 class 102031 non-null object\n",
830
+ " 6 order 102031 non-null object\n",
831
+ " 7 family 102031 non-null object\n",
832
+ " 8 genus 102031 non-null object\n",
833
+ " 9 species 102031 non-null object\n",
834
+ "dtypes: object(10)\n",
835
+ "memory usage: 8.6+ MB\n"
836
+ ]
837
+ }
838
+ ],
839
+ "source": [
840
+ "df_filter = df.loc[df[\"species\"].notna()].copy()\n",
841
+ "df_filter[taxa_cols].info(show_counts=True)"
842
+ ]
843
+ },
844
+ {
845
+ "cell_type": "code",
846
+ "execution_count": 16,
847
+ "metadata": {},
848
+ "outputs": [
849
+ {
850
+ "data": {
851
+ "text/plain": [
852
+ "4"
853
+ ]
854
+ },
855
+ "execution_count": 16,
856
+ "metadata": {},
857
+ "output_type": "execute_result"
858
+ }
859
+ ],
860
+ "source": [
861
+ "df_filter[\"subspecies\"].nunique()"
862
+ ]
863
+ },
864
+ {
865
+ "cell_type": "code",
866
+ "execution_count": 17,
867
+ "metadata": {},
868
+ "outputs": [
869
+ {
870
+ "data": {
871
+ "text/plain": [
872
+ "subspecies\n",
873
+ "thamnophis sirtalis sirtalis 31899\n",
874
+ "troglodytes aedon aedon 5934\n",
875
+ "thamnophis saurita saurita 133\n",
876
+ "nerodia sipedon sipedon 121\n",
877
+ "Name: count, dtype: int64"
878
+ ]
879
+ },
880
+ "execution_count": 17,
881
+ "metadata": {},
882
+ "output_type": "execute_result"
883
+ }
884
+ ],
885
+ "source": [
886
+ "df_filter[\"subspecies\"].value_counts()"
887
+ ]
888
+ },
889
+ {
890
+ "cell_type": "markdown",
891
+ "metadata": {},
892
+ "source": [
893
+ "These are the representative subspecies of these species, so I don't think we'll consider the subspecies, especially since it's only labeled for a few. We can check if these species are in here without subspecies designations."
894
+ ]
895
+ },
896
+ {
897
+ "cell_type": "code",
898
+ "execution_count": 19,
899
+ "metadata": {},
900
+ "outputs": [
901
+ {
902
+ "data": {
903
+ "text/plain": [
904
+ "0"
905
+ ]
906
+ },
907
+ "execution_count": 19,
908
+ "metadata": {},
909
+ "output_type": "execute_result"
910
+ }
911
+ ],
912
+ "source": [
913
+ "df_filter.loc[(df_filter[\"species\"].isin([\"thamnophis sirtalis\", \"troglodytes aedon\", \"thamnophis saurita\", \"nerodia sipedon\"])) & (df_filter[\"subspecies\"].isna())].shape[0]"
914
+ ]
915
+ },
916
+ {
917
+ "cell_type": "markdown",
918
+ "metadata": {},
919
+ "source": [
920
+ "Yes, they're the only way these species are represented, so we can remove the subspecies."
921
+ ]
922
+ },
923
+ {
924
+ "cell_type": "markdown",
925
+ "metadata": {},
926
+ "source": [
927
+ "### Remove extra columns\n",
928
+ "\n",
929
+ "Only need `taxa_cols` (Linnean taxonomy + `original_label`, `scientific_name`, and `common_name`) and `id_cols`."
930
+ ]
931
+ },
932
+ {
933
+ "cell_type": "code",
934
+ "execution_count": 29,
935
+ "metadata": {},
936
+ "outputs": [],
937
+ "source": [
938
+ "id_cols = ['dataset_name',\n",
939
+ " 'url_gcp',\n",
940
+ " 'url_aws',\n",
941
+ " 'url_azure',\n",
942
+ " 'image_id',\n",
943
+ " 'sequence_id',\n",
944
+ " 'location_id',\n",
945
+ " 'frame_num']\n",
946
+ "\n",
947
+ "cols_to_keep = [col for col in list(df.columns) if (col in id_cols or col in taxa_cols)]"
948
+ ]
949
+ },
950
+ {
951
+ "cell_type": "code",
952
+ "execution_count": 30,
953
+ "metadata": {},
954
+ "outputs": [
955
+ {
956
+ "data": {
957
+ "text/plain": [
958
+ "['dataset_name',\n",
959
+ " 'url_gcp',\n",
960
+ " 'url_aws',\n",
961
+ " 'url_azure',\n",
962
+ " 'image_id',\n",
963
+ " 'sequence_id',\n",
964
+ " 'location_id',\n",
965
+ " 'frame_num',\n",
966
+ " 'original_label',\n",
967
+ " 'scientific_name',\n",
968
+ " 'common_name',\n",
969
+ " 'kingdom',\n",
970
+ " 'phylum',\n",
971
+ " 'class',\n",
972
+ " 'order',\n",
973
+ " 'family',\n",
974
+ " 'genus',\n",
975
+ " 'species']"
976
+ ]
977
+ },
978
+ "execution_count": 30,
979
+ "metadata": {},
980
+ "output_type": "execute_result"
981
+ }
982
+ ],
983
+ "source": [
984
+ "cols_to_keep"
985
+ ]
986
+ },
987
+ {
988
+ "cell_type": "markdown",
989
+ "metadata": {},
990
+ "source": [
991
+ "Let's add a number of images column (by `scientific_name`)."
992
+ ]
993
+ },
994
+ {
995
+ "cell_type": "code",
996
+ "execution_count": 31,
997
+ "metadata": {},
998
+ "outputs": [
999
+ {
1000
+ "data": {
1001
+ "text/html": [
1002
+ "<div>\n",
1003
+ "<style scoped>\n",
1004
+ " .dataframe tbody tr th:only-of-type {\n",
1005
+ " vertical-align: middle;\n",
1006
+ " }\n",
1007
+ "\n",
1008
+ " .dataframe tbody tr th {\n",
1009
+ " vertical-align: top;\n",
1010
+ " }\n",
1011
+ "\n",
1012
+ " .dataframe thead th {\n",
1013
+ " text-align: right;\n",
1014
+ " }\n",
1015
+ "</style>\n",
1016
+ "<table border=\"1\" class=\"dataframe\">\n",
1017
+ " <thead>\n",
1018
+ " <tr style=\"text-align: right;\">\n",
1019
+ " <th></th>\n",
1020
+ " <th>dataset_name</th>\n",
1021
+ " <th>url_gcp</th>\n",
1022
+ " <th>url_aws</th>\n",
1023
+ " <th>url_azure</th>\n",
1024
+ " <th>image_id</th>\n",
1025
+ " <th>sequence_id</th>\n",
1026
+ " <th>location_id</th>\n",
1027
+ " <th>frame_num</th>\n",
1028
+ " <th>original_label</th>\n",
1029
+ " <th>scientific_name</th>\n",
1030
+ " <th>...</th>\n",
1031
+ " <th>family</th>\n",
1032
+ " <th>subfamily</th>\n",
1033
+ " <th>tribe</th>\n",
1034
+ " <th>genus</th>\n",
1035
+ " <th>species</th>\n",
1036
+ " <th>subspecies</th>\n",
1037
+ " <th>variety</th>\n",
1038
+ " <th>multi_species</th>\n",
1039
+ " <th>num_species</th>\n",
1040
+ " <th>num_sp_images</th>\n",
1041
+ " </tr>\n",
1042
+ " </thead>\n",
1043
+ " <tbody>\n",
1044
+ " <tr>\n",
1045
+ " <th>0</th>\n",
1046
+ " <td>Ohio Small Animals</td>\n",
1047
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1048
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1049
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1050
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1051
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
1052
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
1053
+ " <td>0</td>\n",
1054
+ " <td>american_bullfrog</td>\n",
1055
+ " <td>lithobates catesbeianus</td>\n",
1056
+ " <td>...</td>\n",
1057
+ " <td>ranidae</td>\n",
1058
+ " <td>NaN</td>\n",
1059
+ " <td>NaN</td>\n",
1060
+ " <td>lithobates</td>\n",
1061
+ " <td>lithobates catesbeianus</td>\n",
1062
+ " <td>NaN</td>\n",
1063
+ " <td>NaN</td>\n",
1064
+ " <td>False</td>\n",
1065
+ " <td>1.0</td>\n",
1066
+ " <td>12.0</td>\n",
1067
+ " </tr>\n",
1068
+ " <tr>\n",
1069
+ " <th>1</th>\n",
1070
+ " <td>Ohio Small Animals</td>\n",
1071
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1072
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1073
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1074
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1075
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
1076
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
1077
+ " <td>1</td>\n",
1078
+ " <td>american_bullfrog</td>\n",
1079
+ " <td>lithobates catesbeianus</td>\n",
1080
+ " <td>...</td>\n",
1081
+ " <td>ranidae</td>\n",
1082
+ " <td>NaN</td>\n",
1083
+ " <td>NaN</td>\n",
1084
+ " <td>lithobates</td>\n",
1085
+ " <td>lithobates catesbeianus</td>\n",
1086
+ " <td>NaN</td>\n",
1087
+ " <td>NaN</td>\n",
1088
+ " <td>False</td>\n",
1089
+ " <td>1.0</td>\n",
1090
+ " <td>12.0</td>\n",
1091
+ " </tr>\n",
1092
+ " <tr>\n",
1093
+ " <th>2</th>\n",
1094
+ " <td>Ohio Small Animals</td>\n",
1095
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1096
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1097
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1098
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1099
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
1100
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
1101
+ " <td>2</td>\n",
1102
+ " <td>american_bullfrog</td>\n",
1103
+ " <td>lithobates catesbeianus</td>\n",
1104
+ " <td>...</td>\n",
1105
+ " <td>ranidae</td>\n",
1106
+ " <td>NaN</td>\n",
1107
+ " <td>NaN</td>\n",
1108
+ " <td>lithobates</td>\n",
1109
+ " <td>lithobates catesbeianus</td>\n",
1110
+ " <td>NaN</td>\n",
1111
+ " <td>NaN</td>\n",
1112
+ " <td>False</td>\n",
1113
+ " <td>1.0</td>\n",
1114
+ " <td>12.0</td>\n",
1115
+ " </tr>\n",
1116
+ " <tr>\n",
1117
+ " <th>3</th>\n",
1118
+ " <td>Ohio Small Animals</td>\n",
1119
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1120
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1121
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1122
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1123
+ " <td>Ohio Small Animals : location_BIWA4N_sequence_...</td>\n",
1124
+ " <td>Ohio Small Animals : BIWA4N</td>\n",
1125
+ " <td>0</td>\n",
1126
+ " <td>american_bullfrog</td>\n",
1127
+ " <td>lithobates catesbeianus</td>\n",
1128
+ " <td>...</td>\n",
1129
+ " <td>ranidae</td>\n",
1130
+ " <td>NaN</td>\n",
1131
+ " <td>NaN</td>\n",
1132
+ " <td>lithobates</td>\n",
1133
+ " <td>lithobates catesbeianus</td>\n",
1134
+ " <td>NaN</td>\n",
1135
+ " <td>NaN</td>\n",
1136
+ " <td>False</td>\n",
1137
+ " <td>1.0</td>\n",
1138
+ " <td>12.0</td>\n",
1139
+ " </tr>\n",
1140
+ " <tr>\n",
1141
+ " <th>4</th>\n",
1142
+ " <td>Ohio Small Animals</td>\n",
1143
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1144
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1145
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1146
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1147
+ " <td>Ohio Small Animals : location_BIWA4N_sequence_...</td>\n",
1148
+ " <td>Ohio Small Animals : BIWA4N</td>\n",
1149
+ " <td>1</td>\n",
1150
+ " <td>american_bullfrog</td>\n",
1151
+ " <td>lithobates catesbeianus</td>\n",
1152
+ " <td>...</td>\n",
1153
+ " <td>ranidae</td>\n",
1154
+ " <td>NaN</td>\n",
1155
+ " <td>NaN</td>\n",
1156
+ " <td>lithobates</td>\n",
1157
+ " <td>lithobates catesbeianus</td>\n",
1158
+ " <td>NaN</td>\n",
1159
+ " <td>NaN</td>\n",
1160
+ " <td>False</td>\n",
1161
+ " <td>1.0</td>\n",
1162
+ " <td>12.0</td>\n",
1163
+ " </tr>\n",
1164
+ " </tbody>\n",
1165
+ "</table>\n",
1166
+ "<p>5 rows × 35 columns</p>\n",
1167
+ "</div>"
1168
+ ],
1169
+ "text/plain": [
1170
+ " dataset_name url_gcp \\\n",
1171
+ "0 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1172
+ "1 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1173
+ "2 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1174
+ "3 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1175
+ "4 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1176
+ "\n",
1177
+ " url_aws \\\n",
1178
+ "0 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1179
+ "1 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1180
+ "2 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1181
+ "3 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1182
+ "4 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1183
+ "\n",
1184
+ " url_azure \\\n",
1185
+ "0 https://lilawildlife.blob.core.windows.net/lil... \n",
1186
+ "1 https://lilawildlife.blob.core.windows.net/lil... \n",
1187
+ "2 https://lilawildlife.blob.core.windows.net/lil... \n",
1188
+ "3 https://lilawildlife.blob.core.windows.net/lil... \n",
1189
+ "4 https://lilawildlife.blob.core.windows.net/lil... \n",
1190
+ "\n",
1191
+ " image_id \\\n",
1192
+ "0 Ohio Small Animals : Images/Sorted_by_species/... \n",
1193
+ "1 Ohio Small Animals : Images/Sorted_by_species/... \n",
1194
+ "2 Ohio Small Animals : Images/Sorted_by_species/... \n",
1195
+ "3 Ohio Small Animals : Images/Sorted_by_species/... \n",
1196
+ "4 Ohio Small Animals : Images/Sorted_by_species/... \n",
1197
+ "\n",
1198
+ " sequence_id \\\n",
1199
+ "0 Ohio Small Animals : location_BIWA3N_sequence_... \n",
1200
+ "1 Ohio Small Animals : location_BIWA3N_sequence_... \n",
1201
+ "2 Ohio Small Animals : location_BIWA3N_sequence_... \n",
1202
+ "3 Ohio Small Animals : location_BIWA4N_sequence_... \n",
1203
+ "4 Ohio Small Animals : location_BIWA4N_sequence_... \n",
1204
+ "\n",
1205
+ " location_id frame_num original_label \\\n",
1206
+ "0 Ohio Small Animals : BIWA3N 0 american_bullfrog \n",
1207
+ "1 Ohio Small Animals : BIWA3N 1 american_bullfrog \n",
1208
+ "2 Ohio Small Animals : BIWA3N 2 american_bullfrog \n",
1209
+ "3 Ohio Small Animals : BIWA4N 0 american_bullfrog \n",
1210
+ "4 Ohio Small Animals : BIWA4N 1 american_bullfrog \n",
1211
+ "\n",
1212
+ " scientific_name ... family subfamily tribe genus \\\n",
1213
+ "0 lithobates catesbeianus ... ranidae NaN NaN lithobates \n",
1214
+ "1 lithobates catesbeianus ... ranidae NaN NaN lithobates \n",
1215
+ "2 lithobates catesbeianus ... ranidae NaN NaN lithobates \n",
1216
+ "3 lithobates catesbeianus ... ranidae NaN NaN lithobates \n",
1217
+ "4 lithobates catesbeianus ... ranidae NaN NaN lithobates \n",
1218
+ "\n",
1219
+ " species subspecies variety multi_species num_species \\\n",
1220
+ "0 lithobates catesbeianus NaN NaN False 1.0 \n",
1221
+ "1 lithobates catesbeianus NaN NaN False 1.0 \n",
1222
+ "2 lithobates catesbeianus NaN NaN False 1.0 \n",
1223
+ "3 lithobates catesbeianus NaN NaN False 1.0 \n",
1224
+ "4 lithobates catesbeianus NaN NaN False 1.0 \n",
1225
+ "\n",
1226
+ " num_sp_images \n",
1227
+ "0 12.0 \n",
1228
+ "1 12.0 \n",
1229
+ "2 12.0 \n",
1230
+ "3 12.0 \n",
1231
+ "4 12.0 \n",
1232
+ "\n",
1233
+ "[5 rows x 35 columns]"
1234
+ ]
1235
+ },
1236
+ "execution_count": 31,
1237
+ "metadata": {},
1238
+ "output_type": "execute_result"
1239
+ }
1240
+ ],
1241
+ "source": [
1242
+ "for sci_name in list(df_filter[\"scientific_name\"].unique()):\n",
1243
+ " df_filter.loc[df_filter[\"scientific_name\"] == sci_name, \"num_sp_images\"] = df_filter.loc[df_filter[\"scientific_name\"] == sci_name].shape[0]\n",
1244
+ "\n",
1245
+ "df_filter.head()"
1246
+ ]
1247
+ },
1248
+ {
1249
+ "cell_type": "code",
1250
+ "execution_count": 32,
1251
+ "metadata": {},
1252
+ "outputs": [],
1253
+ "source": [
1254
+ "cols_to_keep.append(\"num_sp_images\")"
1255
+ ]
1256
+ },
1257
+ {
1258
+ "cell_type": "code",
1259
+ "execution_count": 33,
1260
+ "metadata": {},
1261
+ "outputs": [
1262
+ {
1263
+ "data": {
1264
+ "text/html": [
1265
+ "<div>\n",
1266
+ "<style scoped>\n",
1267
+ " .dataframe tbody tr th:only-of-type {\n",
1268
+ " vertical-align: middle;\n",
1269
+ " }\n",
1270
+ "\n",
1271
+ " .dataframe tbody tr th {\n",
1272
+ " vertical-align: top;\n",
1273
+ " }\n",
1274
+ "\n",
1275
+ " .dataframe thead th {\n",
1276
+ " text-align: right;\n",
1277
+ " }\n",
1278
+ "</style>\n",
1279
+ "<table border=\"1\" class=\"dataframe\">\n",
1280
+ " <thead>\n",
1281
+ " <tr style=\"text-align: right;\">\n",
1282
+ " <th></th>\n",
1283
+ " <th>dataset_name</th>\n",
1284
+ " <th>url_gcp</th>\n",
1285
+ " <th>url_aws</th>\n",
1286
+ " <th>url_azure</th>\n",
1287
+ " <th>image_id</th>\n",
1288
+ " <th>sequence_id</th>\n",
1289
+ " <th>location_id</th>\n",
1290
+ " <th>frame_num</th>\n",
1291
+ " <th>original_label</th>\n",
1292
+ " <th>scientific_name</th>\n",
1293
+ " <th>common_name</th>\n",
1294
+ " <th>kingdom</th>\n",
1295
+ " <th>phylum</th>\n",
1296
+ " <th>class</th>\n",
1297
+ " <th>order</th>\n",
1298
+ " <th>family</th>\n",
1299
+ " <th>genus</th>\n",
1300
+ " <th>species</th>\n",
1301
+ " <th>num_sp_images</th>\n",
1302
+ " </tr>\n",
1303
+ " </thead>\n",
1304
+ " <tbody>\n",
1305
+ " <tr>\n",
1306
+ " <th>0</th>\n",
1307
+ " <td>Ohio Small Animals</td>\n",
1308
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1309
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1310
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1311
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1312
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
1313
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
1314
+ " <td>0</td>\n",
1315
+ " <td>american_bullfrog</td>\n",
1316
+ " <td>lithobates catesbeianus</td>\n",
1317
+ " <td>american bullfrog</td>\n",
1318
+ " <td>animalia</td>\n",
1319
+ " <td>chordata</td>\n",
1320
+ " <td>amphibia</td>\n",
1321
+ " <td>anura</td>\n",
1322
+ " <td>ranidae</td>\n",
1323
+ " <td>lithobates</td>\n",
1324
+ " <td>lithobates catesbeianus</td>\n",
1325
+ " <td>12.0</td>\n",
1326
+ " </tr>\n",
1327
+ " <tr>\n",
1328
+ " <th>1</th>\n",
1329
+ " <td>Ohio Small Animals</td>\n",
1330
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1331
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1332
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1333
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1334
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
1335
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
1336
+ " <td>1</td>\n",
1337
+ " <td>american_bullfrog</td>\n",
1338
+ " <td>lithobates catesbeianus</td>\n",
1339
+ " <td>american bullfrog</td>\n",
1340
+ " <td>animalia</td>\n",
1341
+ " <td>chordata</td>\n",
1342
+ " <td>amphibia</td>\n",
1343
+ " <td>anura</td>\n",
1344
+ " <td>ranidae</td>\n",
1345
+ " <td>lithobates</td>\n",
1346
+ " <td>lithobates catesbeianus</td>\n",
1347
+ " <td>12.0</td>\n",
1348
+ " </tr>\n",
1349
+ " <tr>\n",
1350
+ " <th>2</th>\n",
1351
+ " <td>Ohio Small Animals</td>\n",
1352
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1353
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1354
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1355
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1356
+ " <td>Ohio Small Animals : location_BIWA3N_sequence_...</td>\n",
1357
+ " <td>Ohio Small Animals : BIWA3N</td>\n",
1358
+ " <td>2</td>\n",
1359
+ " <td>american_bullfrog</td>\n",
1360
+ " <td>lithobates catesbeianus</td>\n",
1361
+ " <td>american bullfrog</td>\n",
1362
+ " <td>animalia</td>\n",
1363
+ " <td>chordata</td>\n",
1364
+ " <td>amphibia</td>\n",
1365
+ " <td>anura</td>\n",
1366
+ " <td>ranidae</td>\n",
1367
+ " <td>lithobates</td>\n",
1368
+ " <td>lithobates catesbeianus</td>\n",
1369
+ " <td>12.0</td>\n",
1370
+ " </tr>\n",
1371
+ " <tr>\n",
1372
+ " <th>3</th>\n",
1373
+ " <td>Ohio Small Animals</td>\n",
1374
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1375
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1376
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1377
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1378
+ " <td>Ohio Small Animals : location_BIWA4N_sequence_...</td>\n",
1379
+ " <td>Ohio Small Animals : BIWA4N</td>\n",
1380
+ " <td>0</td>\n",
1381
+ " <td>american_bullfrog</td>\n",
1382
+ " <td>lithobates catesbeianus</td>\n",
1383
+ " <td>american bullfrog</td>\n",
1384
+ " <td>animalia</td>\n",
1385
+ " <td>chordata</td>\n",
1386
+ " <td>amphibia</td>\n",
1387
+ " <td>anura</td>\n",
1388
+ " <td>ranidae</td>\n",
1389
+ " <td>lithobates</td>\n",
1390
+ " <td>lithobates catesbeianus</td>\n",
1391
+ " <td>12.0</td>\n",
1392
+ " </tr>\n",
1393
+ " <tr>\n",
1394
+ " <th>4</th>\n",
1395
+ " <td>Ohio Small Animals</td>\n",
1396
+ " <td>https://storage.googleapis.com/public-datasets...</td>\n",
1397
+ " <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
1398
+ " <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
1399
+ " <td>Ohio Small Animals : Images/Sorted_by_species/...</td>\n",
1400
+ " <td>Ohio Small Animals : location_BIWA4N_sequence_...</td>\n",
1401
+ " <td>Ohio Small Animals : BIWA4N</td>\n",
1402
+ " <td>1</td>\n",
1403
+ " <td>american_bullfrog</td>\n",
1404
+ " <td>lithobates catesbeianus</td>\n",
1405
+ " <td>american bullfrog</td>\n",
1406
+ " <td>animalia</td>\n",
1407
+ " <td>chordata</td>\n",
1408
+ " <td>amphibia</td>\n",
1409
+ " <td>anura</td>\n",
1410
+ " <td>ranidae</td>\n",
1411
+ " <td>lithobates</td>\n",
1412
+ " <td>lithobates catesbeianus</td>\n",
1413
+ " <td>12.0</td>\n",
1414
+ " </tr>\n",
1415
+ " </tbody>\n",
1416
+ "</table>\n",
1417
+ "</div>"
1418
+ ],
1419
+ "text/plain": [
1420
+ " dataset_name url_gcp \\\n",
1421
+ "0 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1422
+ "1 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1423
+ "2 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1424
+ "3 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1425
+ "4 Ohio Small Animals https://storage.googleapis.com/public-datasets... \n",
1426
+ "\n",
1427
+ " url_aws \\\n",
1428
+ "0 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1429
+ "1 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1430
+ "2 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1431
+ "3 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1432
+ "4 http://us-west-2.opendata.source.coop.s3.amazo... \n",
1433
+ "\n",
1434
+ " url_azure \\\n",
1435
+ "0 https://lilawildlife.blob.core.windows.net/lil... \n",
1436
+ "1 https://lilawildlife.blob.core.windows.net/lil... \n",
1437
+ "2 https://lilawildlife.blob.core.windows.net/lil... \n",
1438
+ "3 https://lilawildlife.blob.core.windows.net/lil... \n",
1439
+ "4 https://lilawildlife.blob.core.windows.net/lil... \n",
1440
+ "\n",
1441
+ " image_id \\\n",
1442
+ "0 Ohio Small Animals : Images/Sorted_by_species/... \n",
1443
+ "1 Ohio Small Animals : Images/Sorted_by_species/... \n",
1444
+ "2 Ohio Small Animals : Images/Sorted_by_species/... \n",
1445
+ "3 Ohio Small Animals : Images/Sorted_by_species/... \n",
1446
+ "4 Ohio Small Animals : Images/Sorted_by_species/... \n",
1447
+ "\n",
1448
+ " sequence_id \\\n",
1449
+ "0 Ohio Small Animals : location_BIWA3N_sequence_... \n",
1450
+ "1 Ohio Small Animals : location_BIWA3N_sequence_... \n",
1451
+ "2 Ohio Small Animals : location_BIWA3N_sequence_... \n",
1452
+ "3 Ohio Small Animals : location_BIWA4N_sequence_... \n",
1453
+ "4 Ohio Small Animals : location_BIWA4N_sequence_... \n",
1454
+ "\n",
1455
+ " location_id frame_num original_label \\\n",
1456
+ "0 Ohio Small Animals : BIWA3N 0 american_bullfrog \n",
1457
+ "1 Ohio Small Animals : BIWA3N 1 american_bullfrog \n",
1458
+ "2 Ohio Small Animals : BIWA3N 2 american_bullfrog \n",
1459
+ "3 Ohio Small Animals : BIWA4N 0 american_bullfrog \n",
1460
+ "4 Ohio Small Animals : BIWA4N 1 american_bullfrog \n",
1461
+ "\n",
1462
+ " scientific_name common_name kingdom phylum class \\\n",
1463
+ "0 lithobates catesbeianus american bullfrog animalia chordata amphibia \n",
1464
+ "1 lithobates catesbeianus american bullfrog animalia chordata amphibia \n",
1465
+ "2 lithobates catesbeianus american bullfrog animalia chordata amphibia \n",
1466
+ "3 lithobates catesbeianus american bullfrog animalia chordata amphibia \n",
1467
+ "4 lithobates catesbeianus american bullfrog animalia chordata amphibia \n",
1468
+ "\n",
1469
+ " order family genus species num_sp_images \n",
1470
+ "0 anura ranidae lithobates lithobates catesbeianus 12.0 \n",
1471
+ "1 anura ranidae lithobates lithobates catesbeianus 12.0 \n",
1472
+ "2 anura ranidae lithobates lithobates catesbeianus 12.0 \n",
1473
+ "3 anura ranidae lithobates lithobates catesbeianus 12.0 \n",
1474
+ "4 anura ranidae lithobates lithobates catesbeianus 12.0 "
1475
+ ]
1476
+ },
1477
+ "execution_count": 33,
1478
+ "metadata": {},
1479
+ "output_type": "execute_result"
1480
+ }
1481
+ ],
1482
+ "source": [
1483
+ "df_reduced = df_filter[cols_to_keep].copy()\n",
1484
+ "df_reduced.head()"
1485
+ ]
1486
+ },
1487
+ {
1488
+ "cell_type": "markdown",
1489
+ "metadata": {},
1490
+ "source": [
1491
+ "### Reduce to no more than 10K images per species"
1492
+ ]
1493
+ },
1494
+ {
1495
+ "cell_type": "code",
1496
+ "execution_count": 35,
1497
+ "metadata": {},
1498
+ "outputs": [
1499
+ {
1500
+ "name": "stdout",
1501
+ "output_type": "stream",
1502
+ "text": [
1503
+ "30848\n"
1504
+ ]
1505
+ }
1506
+ ],
1507
+ "source": [
1508
+ "imgs_to_keep = list(df_reduced.loc[df_reduced[\"num_sp_images\"] <= 10000, \"image_id\"])\n",
1509
+ "print(len(imgs_to_keep))"
1510
+ ]
1511
+ },
1512
+ {
1513
+ "cell_type": "code",
1514
+ "execution_count": 38,
1515
+ "metadata": {},
1516
+ "outputs": [
1517
+ {
1518
+ "data": {
1519
+ "text/plain": [
1520
+ "4"
1521
+ ]
1522
+ },
1523
+ "execution_count": 38,
1524
+ "metadata": {},
1525
+ "output_type": "execute_result"
1526
+ }
1527
+ ],
1528
+ "source": [
1529
+ "high_num_classes = list(df_reduced.loc[~df_reduced[\"image_id\"].isin(imgs_to_keep), \"scientific_name\"].unique())\n",
1530
+ "len(high_num_classes)"
1531
+ ]
1532
+ },
1533
+ {
1534
+ "cell_type": "code",
1535
+ "execution_count": 39,
1536
+ "metadata": {},
1537
+ "outputs": [
1538
+ {
1539
+ "data": {
1540
+ "text/plain": [
1541
+ "70848"
1542
+ ]
1543
+ },
1544
+ "execution_count": 39,
1545
+ "metadata": {},
1546
+ "output_type": "execute_result"
1547
+ }
1548
+ ],
1549
+ "source": [
1550
+ "for sci_name in high_num_classes:\n",
1551
+ " sample_set = list(df_reduced.loc[df_reduced[\"scientific_name\"] == sci_name].sample(10000, random_state = 614)[\"image_id\"])\n",
1552
+ " imgs_to_keep = imgs_to_keep + sample_set\n",
1553
+ "\n",
1554
+ "len(imgs_to_keep)"
1555
+ ]
1556
+ },
1557
+ {
1558
+ "cell_type": "code",
1559
+ "execution_count": 40,
1560
+ "metadata": {},
1561
+ "outputs": [],
1562
+ "source": [
1563
+ "df_reduced_upper_bound = df_reduced.loc[df_reduced[\"image_id\"].isin(imgs_to_keep)].copy()\n",
1564
+ "df_reduced_upper_bound.to_csv(\"../data/potential-test-sets/filtered/ohio-small-animals-upper-bound.csv\", index = False)"
1565
+ ]
1566
+ },
1567
+ {
1568
+ "cell_type": "code",
1569
+ "execution_count": 41,
1570
+ "metadata": {},
1571
+ "outputs": [],
1572
+ "source": [
1573
+ "df_reduced_upper_lower = df_reduced_upper_bound.loc[df_reduced_upper_bound[\"num_sp_images\"] >= 10].copy()\n",
1574
+ "df_reduced_upper_lower.to_csv(\"../data/potential-test-sets/filtered/ohio-small-animals-upper-lower-bound.csv\", index = False)"
1575
+ ]
1576
+ },
1577
+ {
1578
+ "cell_type": "markdown",
1579
+ "metadata": {},
1580
+ "source": [
1581
+ "### Randomly sample to balanced set (12 images per species)\n"
1582
+ ]
1583
+ },
1584
+ {
1585
+ "cell_type": "code",
1586
+ "execution_count": 42,
1587
+ "metadata": {},
1588
+ "outputs": [
1589
+ {
1590
+ "data": {
1591
+ "text/plain": [
1592
+ "468"
1593
+ ]
1594
+ },
1595
+ "execution_count": 42,
1596
+ "metadata": {},
1597
+ "output_type": "execute_result"
1598
+ }
1599
+ ],
1600
+ "source": [
1601
+ "balanced_set = []\n",
1602
+ "for sci_name in list(df_reduced[\"scientific_name\"].unique()):\n",
1603
+ " temp = df_reduced.loc[df_reduced[\"scientific_name\"] == sci_name].copy()\n",
1604
+ " if temp.shape[0] < 12:\n",
1605
+ " continue\n",
1606
+ " sample_set = list(temp.sample(12, random_state = 614)[\"image_id\"])\n",
1607
+ " balanced_set = balanced_set + sample_set\n",
1608
+ "\n",
1609
+ "len(balanced_set)"
1610
+ ]
1611
+ },
1612
+ {
1613
+ "cell_type": "markdown",
1614
+ "metadata": {},
1615
+ "source": [
1616
+ "Filter to just balanced set and drop the number of species column since it's been balanced to 12 each."
1617
+ ]
1618
+ },
1619
+ {
1620
+ "cell_type": "code",
1621
+ "execution_count": 43,
1622
+ "metadata": {},
1623
+ "outputs": [],
1624
+ "source": [
1625
+ "df_balanced = df_reduced.loc[df_reduced[\"image_id\"].isin(balanced_set)].copy()\n",
1626
+ "df_balanced.drop(columns = [\"num_sp_images\"], inplace = True)\n",
1627
+ "df_balanced.to_csv(\"../data/potential-test-sets/filtered/ohio-small-animals-balanced.csv\", index = False)"
1628
+ ]
1629
+ },
1630
+ {
1631
+ "cell_type": "code",
1632
+ "execution_count": null,
1633
+ "metadata": {},
1634
+ "outputs": [],
1635
+ "source": []
1636
+ }
1637
+ ],
1638
+ "metadata": {
1639
+ "jupytext": {
1640
+ "formats": "ipynb,py:percent"
1641
+ },
1642
+ "kernelspec": {
1643
+ "display_name": "data-dev",
1644
+ "language": "python",
1645
+ "name": "python3"
1646
+ },
1647
+ "language_info": {
1648
+ "codemirror_mode": {
1649
+ "name": "ipython",
1650
+ "version": 3
1651
+ },
1652
+ "file_extension": ".py",
1653
+ "mimetype": "text/x-python",
1654
+ "name": "python",
1655
+ "nbconvert_exporter": "python",
1656
+ "pygments_lexer": "ipython3",
1657
+ "version": "3.12.3"
1658
+ }
1659
+ },
1660
+ "nbformat": 4,
1661
+ "nbformat_minor": 2
1662
+ }
notebooks/lilabc_test-orinoquia.ipynb ADDED
The diff for this file is too large to render. See raw diff