Datasets:
File size: 4,148 Bytes
c528b1f 3c3abf4 c528b1f 7c045cd 8144255 c528b1f 3c3abf4 e6a5b9a 3c3abf4 7c045cd 3c3abf4 d3e75e8 3c3abf4 8666ba0 3c3abf4 79f694b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 | ---
pretty_name: DisImpact
tags:
- disaster
- social-media
- hurricane
- wildfire
- reddit
- tiktok
- youtube
- disaster-impact
size_categories:
- 10K<n<100K
license: cc-by-4.0
language:
- en
dataset_info:
features:
- name: id
dtype: string
- name: label
dtype: int64
---
# DisImpact: Disaster Physi-Social Impact Classification
## Dataset Description
Dataset of the paper "[DisImpact: Quantifying the Physi-Social Impact of Natural Disasters Through Social Media
](https://arxiv.org/pdf/2605.20646v1)". This dataset contains post IDs and disaster impact labels for **hurricane** and **wildfire** across three platforms: **Reddit**, **TikTok**, and **YouTube**.
## Data Format
Each disaster–platform subset is stored in a CSV file named `<disaster>_<platform>_social.csv`, with two columns:
| Column | Type | Description |
|---|---|---|
| `id` | string | id of the post. |
| `label` | integer | Disaster impact category, numbered from 1 to 11. |
**Always read `id` as a string.** Numeric conversion can alter identifiers, especially long TikTok IDs. CSV files do not preserve column types.
## Label Definitions
Labels 1–10 follow the category order in [Table 3 of the paper](https://arxiv.org/html/2605.20646v1#S3.T3). Descriptions below are abbreviated.
| Label | Domain | Category | Description |
|---|---|---|---|
| 1 | Physical | Casualties & Injuries (CINJ) | Deaths, injuries, missing people or animals, and immediate medical emergencies. |
| 2 | Physical | Evacuations & Displacement (EVAC) | Forced evacuation, displacement, or relocation to shelters. |
| 3 | Physical | Infrastructure & Utility Damage (INFR) | Damaged infrastructure or disrupted utilities. |
| 4 | Physical | Environmental Damage (ENVD) | Ecosystem damage, agricultural losses, or environmental contamination. |
| 5 | Physical | Resource Shortages (RSRC) | Requests for essential survival supplies. |
| 6 | Social | Public Health (PUBH) | Disease outbreaks, interrupted healthcare, or medical supply shortages. |
| 7 | Social | Emotional & Psychological Distress (EMOT) | Disaster-related trauma, anxiety, grief, or other emotional suffering. |
| 8 | Social | Bias Narratives (BIAS) | Discrimination, political blame, or unequal impacts across groups. |
| 9 | Social | Assistance & Recovery (ASST) | Aid, rebuilding, community support, and adaptation. |
| 10 | Social | Socioeconomic Disruption (SECO) | Disruption to livelihoods, housing, education, or economic activity. |
| 11 | Other | None of the Above (OTHER) | None of categories 1–10 applies. |
Label 11 does not necessarily mean that a post is unrelated to the disaster.
## Privacy and Content Access
To protect user privacy, only post IDs are released. We recommend retrieving the full post content via the official APIs of each platform, in accordance with their respective terms of service.
- [Reddit API](https://www.reddit.com/dev/api)
- [TikTok API](https://developers.tiktok.com/products/research-api)
- [YouTube API](https://developers.google.com/youtube/v3)
The released CSV files contain post IDs and classification labels; they do not include post text, images, videos, or user profiles. Content availability may change if posts are deleted, made private, or otherwise become inaccessible.
## Limitations
Labels are model-generated annotations and may contain errors. Each post receives a single category, even when it describes multiple impacts. Social media observations should not be interpreted as a representative measure of all affected populations or as verified disaster incident reports.
## Citation
@inproceedings{yao2026disimpact,
title={DisImpact: Quantifying the Physi-Social Impact of Natural Disasters Through Social Media},
author={Yao, Ruichen and Dasari, Tejna and Meng, Xuanyu and Cao, Elliot and Li, Zelin and Liu, Yifan and Liu, Yaokun and Shang, Lanyu and Wang, Dong},
booktitle={Proceedings of the International AAAI Conference on Web and Social Media},
volume={20},
number={1},
pages={2607--2622},
year={2026}
}
Dataset DOI: [10.57967/hf/10879](https://doi.org/10.57967/hf/10879) |