Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
10K - 100K
License:
updated data-card
Browse files
README.md
CHANGED
|
@@ -1,4 +1,25 @@
|
|
| 1 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
|
| 3 |
## Table of Contents
|
| 4 |
- [Dataset Description](#dataset-description)
|
|
@@ -47,7 +68,7 @@ The HumAID Twitter dataset consists of several thousands of manually annotated t
|
|
| 47 |
- Rescue volunteering or donation effort
|
| 48 |
- Sympathy and support
|
| 49 |
|
| 50 |
-
The resulting annotated dataset consists of 11 labels.
|
| 51 |
|
| 52 |
### Supported Tasks and Benchmark
|
| 53 |
The dataset can be used to train a model for multiclass tweet classification for disaster response. The benchmark results can be found in https://ojs.aaai.org/index.php/ICWSM/article/view/18116/17919.
|
|
@@ -66,7 +87,7 @@ English
|
|
| 66 |
```
|
| 67 |
{
|
| 68 |
|
| 69 |
-
"tweet_text": "@RT_com: URGENT: Death toll in #Ecuador #quake rises to 233 \u2013 President #Correa #1 in #Pakistan",
|
| 70 |
|
| 71 |
"class_label": "injured_or_dead_people"
|
| 72 |
|
|
@@ -92,14 +113,14 @@ English
|
|
| 92 |
|
| 93 |
#### Initial Data Collection and Normalization
|
| 94 |
|
| 95 |
-
Tweets has been collected during several disaster events.
|
| 96 |
|
| 97 |
|
| 98 |
### Annotations
|
| 99 |
|
| 100 |
#### Annotation process
|
| 101 |
|
| 102 |
-
AMT has been used to annotate the dataset. Please check the paper for a more detail.
|
| 103 |
|
| 104 |
#### Who are the annotators?
|
| 105 |
|
|
@@ -124,7 +145,7 @@ AMT has been used to annotate the dataset. Please check the paper for a more det
|
|
| 124 |
|
| 125 |
### Dataset Curators
|
| 126 |
|
| 127 |
-
Authors of the paper.
|
| 128 |
|
| 129 |
|
| 130 |
### Licensing Information
|
|
@@ -144,4 +165,4 @@ Year = {2021},
|
|
| 144 |
publisher={AAAI},
|
| 145 |
address={Online},
|
| 146 |
}
|
| 147 |
-
```
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-sa-4.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- text-classification
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
tags:
|
| 8 |
+
- Disaster
|
| 9 |
+
- Crisis Informatics
|
| 10 |
+
pretty_name: 'HumAID: Human-Annotated Disaster Incidents Data from Twitter'
|
| 11 |
+
size_categories:
|
| 12 |
+
- 10K<n<100K
|
| 13 |
+
dataset_info:
|
| 14 |
+
splits:
|
| 15 |
+
- name: train
|
| 16 |
+
num_examples: 53531
|
| 17 |
+
- name: dev
|
| 18 |
+
num_examples: 7793
|
| 19 |
+
- name: test
|
| 20 |
+
num_examples: 15160
|
| 21 |
+
---
|
| 22 |
+
# HumAID: Human-Annotated Disaster Incidents Data from Twitter
|
| 23 |
|
| 24 |
## Table of Contents
|
| 25 |
- [Dataset Description](#dataset-description)
|
|
|
|
| 68 |
- Rescue volunteering or donation effort
|
| 69 |
- Sympathy and support
|
| 70 |
|
| 71 |
+
The resulting annotated dataset consists of 11 labels.
|
| 72 |
|
| 73 |
### Supported Tasks and Benchmark
|
| 74 |
The dataset can be used to train a model for multiclass tweet classification for disaster response. The benchmark results can be found in https://ojs.aaai.org/index.php/ICWSM/article/view/18116/17919.
|
|
|
|
| 87 |
```
|
| 88 |
{
|
| 89 |
|
| 90 |
+
"tweet_text": "@RT_com: URGENT: Death toll in #Ecuador #quake rises to 233 \u2013 President #Correa #1 in #Pakistan",
|
| 91 |
|
| 92 |
"class_label": "injured_or_dead_people"
|
| 93 |
|
|
|
|
| 113 |
|
| 114 |
#### Initial Data Collection and Normalization
|
| 115 |
|
| 116 |
+
Tweets has been collected during several disaster events.
|
| 117 |
|
| 118 |
|
| 119 |
### Annotations
|
| 120 |
|
| 121 |
#### Annotation process
|
| 122 |
|
| 123 |
+
AMT has been used to annotate the dataset. Please check the paper for a more detail.
|
| 124 |
|
| 125 |
#### Who are the annotators?
|
| 126 |
|
|
|
|
| 145 |
|
| 146 |
### Dataset Curators
|
| 147 |
|
| 148 |
+
Authors of the paper.
|
| 149 |
|
| 150 |
|
| 151 |
### Licensing Information
|
|
|
|
| 165 |
publisher={AAAI},
|
| 166 |
address={Online},
|
| 167 |
}
|
| 168 |
+
```
|