QCRI/HumAID-events

---
license: cc-by-nc-sa-4.0
task_categories:
- text-classification
language:
- en
tags:
- Disaster
- Crisis Informatics
pretty_name: 'HumAID: Human-Annotated Disaster Incidents Data from Twitter -- Event wise dataset'
size_categories:
- 10K<n<100K
dataset_info:
- config_name: hurricane_florence_2018
splits:
- name: train
num_examples: 4384
- name: dev
num_examples: 639
- name: test
num_examples: 1241
- config_name: kaikoura_earthquake_2016
splits:
- name: train
num_examples: 1536
- name: dev
num_examples: 224
- name: test
num_examples: 435
- config_name: kerala_floods_2018
splits:
- name: train
num_examples: 5588
- name: dev
num_examples: 814
- name: test
num_examples: 1582
- config_name: hurricane_harvey_2017
splits:
- name: train
num_examples: 6378
- name: dev
num_examples: 929
- name: test
num_examples: 1805
- config_name: hurricane_maria_2017
splits:
- name: train
num_examples: 5094
- name: dev
num_examples: 742
- name: test
num_examples: 1442
- config_name: midwestern_us_floods_2019
splits:
- name: train
num_examples: 1316
- name: dev
num_examples: 191
- name: test
num_examples: 373
- config_name: puebla_mexico_earthquake_2017
splits:
- name: train
num_examples: 1410
- name: dev
num_examples: 205
- name: test
num_examples: 400
- config_name: maryland_floods_2018
splits:
- name: train
num_examples: 519
- name: dev
num_examples: 75
- name: test
num_examples: 148
- config_name: hurricane_irma_2017
splits:
- name: train
num_examples: 6579
- name: dev
num_examples: 958
- name: test
num_examples: 1862
- config_name: ecuador_earthquake_2016
splits:
- name: train
num_examples: 1094
- name: dev
num_examples: 159
- name: test
num_examples: 310
- config_name: cyclone_idai_2019
splits:
- name: train
num_examples: 2753
- name: dev
num_examples: 401
- name: test
num_examples: 779
- config_name: canada_wildfires_2016
splits:
- name: train
num_examples: 1569
- name: dev
num_examples: 228
- name: test
num_examples: 445
- config_name: italy_earthquake_aug_2016
splits:
- name: train
num_examples: 840
- name: dev
num_examples: 122
- name: test
num_examples: 239
- config_name: greece_wildfires_2018
splits:
- name: train
num_examples: 1060
- name: dev
num_examples: 154
- name: test
num_examples: 301
- config_name: hurricane_dorian_2019
splits:
- name: train
num_examples: 5329
- name: dev
num_examples: 776
- name: test
num_examples: 1508
- config_name: .git
splits:
- name: train
num_examples: 0
- name: dev
num_examples: 0
- name: test
num_examples: 0
- config_name: california_wildfires_2018
splits:
- name: train
num_examples: 5163
- name: dev
num_examples: 752
- name: test
num_examples: 1461
- config_name: pakistan_earthquake_2019
splits:
- name: train
num_examples: 1370
- name: dev
num_examples: 199
- name: test
num_examples: 389
- config_name: hurricane_matthew_2016
splits:
- name: train
num_examples: 1157
- name: dev
num_examples: 168
- name: test
num_examples: 329
- config_name: srilanka_floods_2017
splits:
- name: train
num_examples: 392
- name: dev
num_examples: 57
- name: test
num_examples: 111
configs:
- config_name: hurricane_florence_2018
data_files:
- split: train
path: hurricane_florence_2018/train.json
- split: dev
path: hurricane_florence_2018/dev.json
- split: test
path: hurricane_florence_2018/test.json
- config_name: kaikoura_earthquake_2016
data_files:
- split: train
path: kaikoura_earthquake_2016/train.json
- split: dev
path: kaikoura_earthquake_2016/dev.json
- split: test
path: kaikoura_earthquake_2016/test.json
- config_name: kerala_floods_2018
data_files:
- split: train
path: kerala_floods_2018/train.json
- split: dev
path: kerala_floods_2018/dev.json
- split: test
path: kerala_floods_2018/test.json
- config_name: hurricane_harvey_2017
data_files:
- split: train
path: hurricane_harvey_2017/train.json
- split: dev
path: hurricane_harvey_2017/dev.json
- split: test
path: hurricane_harvey_2017/test.json
- config_name: hurricane_maria_2017
data_files:
- split: train
path: hurricane_maria_2017/train.json
- split: dev
path: hurricane_maria_2017/dev.json
- split: test
path: hurricane_maria_2017/test.json
- config_name: midwestern_us_floods_2019
data_files:
- split: train
path: midwestern_us_floods_2019/train.json
- split: dev
path: midwestern_us_floods_2019/dev.json
- split: test
path: midwestern_us_floods_2019/test.json
- config_name: puebla_mexico_earthquake_2017
data_files:
- split: train
path: puebla_mexico_earthquake_2017/train.json
- split: dev
path: puebla_mexico_earthquake_2017/dev.json
- split: test
path: puebla_mexico_earthquake_2017/test.json
- config_name: maryland_floods_2018
data_files:
- split: train
path: maryland_floods_2018/train.json
- split: dev
path: maryland_floods_2018/dev.json
- split: test
path: maryland_floods_2018/test.json
- config_name: hurricane_irma_2017
data_files:
- split: train
path: hurricane_irma_2017/train.json
- split: dev
path: hurricane_irma_2017/dev.json
- split: test
path: hurricane_irma_2017/test.json
- config_name: ecuador_earthquake_2016
data_files:
- split: train
path: ecuador_earthquake_2016/train.json
- split: dev
path: ecuador_earthquake_2016/dev.json
- split: test
path: ecuador_earthquake_2016/test.json
- config_name: cyclone_idai_2019
data_files:
- split: train
path: cyclone_idai_2019/train.json
- split: dev
path: cyclone_idai_2019/dev.json
- split: test
path: cyclone_idai_2019/test.json
- config_name: canada_wildfires_2016
data_files:
- split: train
path: canada_wildfires_2016/train.json
- split: dev
path: canada_wildfires_2016/dev.json
- split: test
path: canada_wildfires_2016/test.json
- config_name: italy_earthquake_aug_2016
data_files:
- split: train
path: italy_earthquake_aug_2016/train.json
- split: dev
path: italy_earthquake_aug_2016/dev.json
- split: test
path: italy_earthquake_aug_2016/test.json
- config_name: greece_wildfires_2018
data_files:
- split: train
path: greece_wildfires_2018/train.json
- split: dev
path: greece_wildfires_2018/dev.json
- split: test
path: greece_wildfires_2018/test.json
- config_name: hurricane_dorian_2019
data_files:
- split: train
path: hurricane_dorian_2019/train.json
- split: dev
path: hurricane_dorian_2019/dev.json
- split: test
path: hurricane_dorian_2019/test.json
- config_name: california_wildfires_2018
data_files:
- split: train
path: california_wildfires_2018/train.json
- split: dev
path: california_wildfires_2018/dev.json
- split: test
path: california_wildfires_2018/test.json
- config_name: pakistan_earthquake_2019
data_files:
- split: train
path: pakistan_earthquake_2019/train.json
- split: dev
path: pakistan_earthquake_2019/dev.json
- split: test
path: pakistan_earthquake_2019/test.json
- config_name: hurricane_matthew_2016
data_files:
- split: train
path: hurricane_matthew_2016/train.json
- split: dev
path: hurricane_matthew_2016/dev.json
- split: test
path: hurricane_matthew_2016/test.json
- config_name: srilanka_floods_2017
data_files:
- split: train
path: srilanka_floods_2017/train.json
- split: dev
path: srilanka_floods_2017/dev.json
- split: test
path: srilanka_floods_2017/test.json
---
# HumAID: Human-Annotated Disaster Incidents Data from Twitter

## Dataset Description

- **Homepage:** https://crisisnlp.qcri.org/humaid_dataset
- **Repository:** https://crisisnlp.qcri.org/data/humaid/humaid_data_all.zip
- **Paper:** https://ojs.aaai.org/index.php/ICWSM/article/view/18116/17919

### Dataset Summary

The HumAID Twitter dataset consists of several thousands of manually annotated tweets that has been collected during 19 major natural disaster events including earthquakes, hurricanes, wildfires, and floods, which happened from 2016 to 2019 across different parts of the World. The annotations in the provided datasets consists of following humanitarian categories. The dataset consists only english tweets and it is the largest dataset for crisis informatics so far.
** Humanitarian categories **
- Caution and advice
- Displaced people and evacuations
- Dont know cant judge
- Infrastructure and utility damage
- Injured or dead people
- Missing or found people
- Not humanitarian
- Other relevant information
- Requests or urgent needs
- Rescue volunteering or donation effort
- Sympathy and support

The resulting annotated dataset consists of 11 labels.

### Supported Tasks and Benchmark
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.

Dataset is also released with event-wise and JSON objects for further research.
Full set of the dataset can be found in https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/A7NVF7

### Languages

English

## Dataset Structure

### Data Instances

```
{

"tweet_text": "@RT_com: URGENT: Death toll in #Ecuador #quake rises to 233 \u2013 President #Correa #1 in #Pakistan",

"class_label": "injured_or_dead_people"

}
```
### Data Fields

* tweet_text: corresponds to the tweet text.
* class_label: corresponds to a label assigned to a given tweet text

### Data Splits

* Train
* Development
* Test

## Dataset Creation
Tweets has been collected during several disaster events.

### Annotations
AMT has been used to annotate the dataset. Please check the paper for a more detail.

#### Who are the annotators?
- crowdsourced

### Licensing Information

- cc-by-nc-4.0

### Citation Information

```
@inproceedings{humaid2020,
Author = {Firoj Alam, Umair Qazi, Muhammad Imran, Ferda Ofli},
booktitle={Proceedings of the Fifteenth International AAAI Conference on Web and Social Media},
series={ICWSM~'21},
Keywords = {Social Media, Crisis Computing, Tweet Text Classification, Disaster Response},
Title = {HumAID: Human-Annotated Disaster Incidents Data from Twitter},
Year = {2021},
publisher={AAAI},
address={Online},
}
```