File size: 2,935 Bytes
6bb99a3
 
f5ac589
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6bb99a3
f5ac589
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
dc874b3
 
 
 
 
 
 
f5ac589
 
 
dc874b3
 
 
 
 
 
 
 
 
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
90
91
92
93
---
license: cc-by-4.0
task_categories:
- tabular-regression
- tabular-classification
language:
- en
tags:
- disaster-response
- bangladesh
- fairness
- flood
- humanitarian-ai
- equity
- pdna
size_categories:
- n<1K
---

# Bangladesh Flood PDNA 2022

Research datasets from the 2022 Bangladesh floods, curated by
[Farjana Yesmin](https://farjana-yesmin.github.io) for the paper:

> Yesmin, F. & Akter, R. (2026). *Toward Equitable Recovery: A
> Fairness-Aware AI Framework for Prioritizing Post-Flood Aid in
> Bangladesh.* CCAI 2026 (IEEE). [arXiv:2512.22210](https://arxiv.org/abs/2512.22210)

Part of the [FairHealth](https://github.com/Farjana-Yesmin/fairhealth)
library — `pip install fairhealth`

## What Is Included

| File | Description | Rows |
|---|---|---|
| `bangladesh_floods_2022_district_level.csv` | District-level damage and loss | 11 |
| `modeling_dataset_upazila_level.csv` | ML-ready upazila features | 87 |
| `pdna_district_summary.csv` | PDNA official summary by district | 11 |
| `pdna_human_impact.csv` | Human impact indicators | — |
| `pdna_sector_summary.csv` | Sector-level damage breakdown | — |
| `fairness_metrics_summary.csv` | Fair vs baseline model comparison | — |
| `district_performance_comparison.csv` | MAE by district, fair vs baseline | 11 |
| `model_predictions_comparison.csv` | Model prediction outputs | 87 |
| `dataset_metadata.json` | Data provenance and field descriptions | — |
| `data_source_citations.txt` | Full citations for all sources | — |

## Key Findings

The adversarial debiasing model reduces statistical parity difference
by **41.6%** and regional fairness gap by **43.2%** compared to baseline,
with only 2.7 percentage point R² cost (0.784 vs 0.811).

Sunamganj (42.7% poverty, $159.6M damage, 94% inundation) moves from
**rank 14 → rank 6** under the fair model.

## Data Sources

- Ministry of Disaster Management and Relief, Government of Bangladesh.
  *Post Disaster Needs Assessment: Bangladesh Floods 2022* (2023)
- Bangladesh Bureau of Statistics (BBS) — poverty and population data
- World Bank Bangladesh Country Data
- NASA SEDAC — gridded population data
- EM-DAT International Disaster Database

## Usage

```python
from fairhealth.equity.flood_aid import generate_priority_ranking
rankings = generate_priority_ranking(verbose=True)
```

## Citation

```bibtex
@dataset{fairhealth_pdna_2026,
  author    = {Yesmin, Farjana and Akter, Romana},
  title     = {Bangladesh Flood PDNA 2022 Research Dataset},
  year      = {2026},
  publisher = {Hugging Face},
  doi       = {10.57967/hf/8799},
  url       = {https://huggingface.co/datasets/fairhealth/bangladesh-flood-pdna-2022}
}
```

Also cite:
```bibtex
@inproceedings{yesmin2026ccai,
  author = {Yesmin, Farjana and Akter, Romana},
  title  = {Toward Equitable Recovery: A Fairness-Aware AI Framework},
  note   = {CCAI 2026, IEEE, Nanjing. Oral. arXiv:2512.22210},
  year   = {2026}
}
```