electricsheepafrica/african-flood-risk-urban-mapping

---
title: African Flood Risk Urban Mapping
license: cc-by-4.0
language:
- en
tags:
- flood-risk
- urban-planning
- africa
- climate-change
- disaster-risk
- reinsurance
- infrastructure
- climate-adaptation
task_categories:
- tabular-regression
- tabular-classification
pretty_name: African Flood Risk Urban Mapping
dataset_info:
features:
- name: city
dtype: string
- name: country
dtype: string
- name: region
dtype: string
- name: district
dtype: string
- name: scenario
dtype: string
- name: latitude
dtype: float64
- name: longitude
dtype: float64
- name: elevation_m
dtype: float64
- name: coastal
dtype: bool
- name: riverine
dtype: bool
- name: land_use_class
dtype: string
- name: risk_tier
dtype: string
- name: population_density_per_km2
dtype: float64
- name: informality_fraction
dtype: float64
- name: annual_flood_frequency_per_1000
dtype: float64
- name: flood_events_per_year
dtype: float64
- name: return_period_10yr_depth_m
dtype: float64
- name: return_period_25yr_depth_m
dtype: float64
- name: return_period_50yr_depth_m
dtype: float64
- name: return_period_100yr_depth_m
dtype: float64
- name: return_period_200yr_depth_m
dtype: float64
- name: mean_flood_depth_m
dtype: float64
- name: median_flood_depth_m
dtype: float64
- name: p90_flood_depth_m
dtype: float64
- name: max_flood_depth_m
dtype: float64
- name: flood_duration_hours
dtype: float64
- name: district_population
dtype: int64
- name: population_exposed
dtype: int64
- name: population_exposure_fraction
dtype: float64
- name: children_exposed
dtype: int64
- name: elderly_exposed
dtype: int64
- name: infrastructure_vulnerability_index
dtype: float64
- name: drainage_capacity_fraction
dtype: float64
- name: road_network_flood_vulnerability
dtype: float64
- name: power_grid_flood_vulnerability
dtype: float64
- name: water_supply_flood_vulnerability
dtype: float64
- name: healthcare_facility_flood_risk
dtype: float64
- name: lulc_annual_change_rate_pct
dtype: float64
- name: impervious_surface_fraction
dtype: float64
- name: green_space_fraction
dtype: float64
- name: wetland_area_fraction
dtype: float64
- name: ews_coverage_fraction
dtype: float64
- name: ews_lead_time_hours
dtype: float64
- name: community_preparedness_index
dtype: float64
- name: econ_damage_per_capita_per_event_usd
dtype: float64
- name: annual_expected_damage_usd
dtype: float64
- name: annual_expected_damage_pct_of_gdp
dtype: float64
- name: building_replacement_cost_usd
dtype: float64
- name: business_interruption_cost_usd
dtype: float64
- name: building_stock_vulnerability_index
dtype: float64
- name: informal_building_fraction
dtype: float64
- name: flood_resistant_building_fraction
dtype: float64
- name: avg_building_age_years
dtype: float64
- name: multi_story_building_fraction
dtype: float64
- name: social_vulnerability_index
dtype: float64
- name: poverty_rate_fraction
dtype: float64
- name: unemployment_rate_fraction
dtype: float64
- name: access_to_healthcare_index
dtype: float64
- name: education_access_index
dtype: float64
- name: displacement_risk_index
dtype: float64
- name: floodplain_encroachment_rate_pct
dtype: float64
- name: floodplain_area_at_risk_km2
dtype: float64
- name: settlement_in_100yr_floodplain_pct
dtype: float64
- name: composite_flood_risk_score
dtype: float64
- name: flood_risk_category
dtype: string
- name: climate_attribution_factor
dtype: float64
- name: adaptation_readiness_score
dtype: float64
splits:
- name: train
num_bytes: 52428800
num_examples: 23148
download_size: 15728640
dataset_size: 52428800
configs:
- config_name: historical_baseline
description: "Historical baseline conditions (1991-2020 average). Represents current flood risk based on observed historical patterns. Parameters calibrated from World Bank flood risk assessments, AfDB urban flood management reports, and UNDRR GAR 2023."
data_files:
- split: train
path: historical_baseline.csv
- config_name: climate_amplified
description: "Climate-amplified scenario (RCP 6.0 / SSP2-4.5, 2040-2060). Projects flood risk under mid-range climate change with 50-80% increase in extreme precipitation events, sea level rise impacts on coastal cities, and accelerated urbanization. Calibrated from IPCC AR6 Africa Chapter and World Bank climate risk projections."
data_files:
- split: train
path: climate_amplified.csv
- config_name: resilient_infrastructure
description: "Resilient infrastructure scenario with adaptation investments. Models flood risk under comprehensive adaptation including green infrastructure, upgraded drainage systems, flood-resistant construction standards, and expanded early warning systems. Based on AfDB adaptation investment frameworks and World Bank resilient city guidelines."
data_files:
- split: train
path: resilient_infrastructure.csv
---

# African Flood Risk Urban Mapping

## Overview

A comprehensive flood risk dataset covering **18 major African cities** across **216 districts/wards** with **3 climate and infrastructure scenarios**, totaling **23,148 records** with **67 features**. Designed for reinsurers, urban planners, climate researchers, and disaster risk management professionals.

## Cities Covered

| City | Country | Region | Population | Coastal | Riverine |
|------|---------|--------|-----------|---------|----------|
| Lagos | Nigeria | West Africa | 15.4M | Yes | Yes |
| Cairo | Egypt | North Africa | 20.9M | No | Yes |
| Johannesburg | South Africa | Southern Africa | 5.6M | No | No |
| Nairobi | Kenya | East Africa | 4.7M | No | Yes |
| Addis Ababa | Ethiopia | East Africa | 5.2M | No | Yes |
| Accra | Ghana | West Africa | 2.6M | Yes | Yes |
| Dakar | Senegal | West Africa | 3.1M | Yes | No |
| Kinshasa | DR Congo | Central Africa | 15.0M | No | Yes |
| Dar es Salaam | Tanzania | East Africa | 7.4M | Yes | Yes |
| Kampala | Uganda | East Africa | 3.5M | No | Yes |
| Casablanca | Morocco | North Africa | 3.7M | Yes | No |
| Tunis | Tunisia | North Africa | 2.9M | Yes | Yes |
| Luanda | Angola | Southern Africa | 8.3M | Yes | No |
| Maputo | Mozambique | Southern Africa | 1.2M | Yes | Yes |
| Kigali | Rwanda | East Africa | 1.1M | No | Yes |
| Bamako | Mali | West Africa | 2.5M | No | Yes |
| Abidjan | Cote d'Ivoire | West Africa | 5.5M | Yes | Yes |
| Douala | Cameroon | Central Africa | 3.6M | Yes | Yes |

## Scenarios

### historical_baseline
Historical baseline conditions (1991-2020 average). Represents current flood risk based on observed historical patterns. Parameters calibrated from World Bank flood risk assessments, AfDB urban flood management reports, and UNDRR GAR 2023.

### climate_amplified
Climate-amplified scenario (RCP 6.0 / SSP2-4.5, 2040-2060). Projects flood risk under mid-range climate change with 50-80% increase in extreme precipitation events, sea level rise impacts on coastal cities, and accelerated urbanization. Calibrated from IPCC AR6 Africa Chapter and World Bank climate risk projections.

Key multipliers vs baseline:
- Flood frequency: 1.65x
- Flood depth: 1.35x
- Population exposure: 1.25x
- Economic damage: 1.55x
- Return periods shortened (100yr events become ~60yr)

### resilient_infrastructure
Resilient infrastructure scenario with adaptation investments. Models flood risk under comprehensive adaptation including green infrastructure, upgraded drainage systems, flood-resistant construction standards, and expanded early warning systems. Based on AfDB adaptation investment frameworks and World Bank resilient city guidelines.

Key multipliers vs baseline:
- Flood frequency: 0.85x
- Flood depth: 0.72x
- Population exposure: 0.80x
- Economic damage: 0.55x
- EWS coverage: 1.85x
- Return periods extended (100yr events become ~140yr)

## Features (67 columns)

### Geographic & Administrative
- `city`, `country`, `region`, `district`
- `latitude`, `longitude`, `elevation_m`
- `coastal`, `riverine`

### Land Use & Urban Form
- `land_use_class` (10 classes: informal_settlement, formal_residential, commercial, industrial, mixed_use, public_institutional, green_space, wetland, transport_corridor, floodplain_agricultural)
- `risk_tier` (low, medium, high, very_high)
- `population_density_per_km2`, `informality_fraction`

### Flood Frequency & Return Periods
- `annual_flood_frequency_per_1000`, `flood_events_per_year`
- `return_period_10yr_depth_m`, `return_period_25yr_depth_m`, `return_period_50yr_depth_m`, `return_period_100yr_depth_m`, `return_period_200yr_depth_m`

### Flood Depth Distribution
- `mean_flood_depth_m`, `median_flood_depth_m`, `p90_flood_depth_m`, `max_flood_depth_m`
- `flood_duration_hours`

### Population Exposure
- `district_population`, `population_exposed`, `population_exposure_fraction`
- `children_exposed`, `elderly_exposed`

### Infrastructure Vulnerability
- `infrastructure_vulnerability_index`
- `drainage_capacity_fraction`
- `road_network_flood_vulnerability`, `power_grid_flood_vulnerability`, `water_supply_flood_vulnerability`, `healthcare_facility_flood_risk`

### Land Use / Land Cover Changes
- `lulc_annual_change_rate_pct`
- `impervious_surface_fraction`, `green_space_fraction`, `wetland_area_fraction`

### Early Warning Systems
- `ews_coverage_fraction`, `ews_lead_time_hours`, `community_preparedness_index`

### Economic Damage Estimates
- `econ_damage_per_capita_per_event_usd`
- `annual_expected_damage_usd`, `annual_expected_damage_pct_of_gdp`
- `building_replacement_cost_usd`, `business_interruption_cost_usd`

### Building Stock Vulnerability
- `building_stock_vulnerability_index`
- `informal_building_fraction`, `flood_resistant_building_fraction`
- `avg_building_age_years`, `multi_story_building_fraction`

### Social Vulnerability Indices
- `social_vulnerability_index`
- `poverty_rate_fraction`, `unemployment_rate_fraction`
- `access_to_healthcare_index`, `education_access_index`
- `displacement_risk_index`

### Floodplain Encroachment
- `floodplain_encroachment_rate_pct`
- `floodplain_area_at_risk_km2`, `settlement_in_100yr_floodplain_pct`

### Composite Metrics
- `composite_flood_risk_score` (0-100)
- `flood_risk_category` (low/medium/high/very_high)
- `climate_attribution_factor`, `adaptation_readiness_score`

## Literature Sources & Calibration

Parameters are calibrated from the following authoritative sources:

1. **World Bank (2021-2024)**: Flood Risk Assessments for African Cities
2. **African Development Bank (2022-2024)**: Urban Flood Management Reports
3. **UNDRR (2023)**: Global Assessment Report on Disaster Risk Reduction
4. **IPCC AR6 (2022)**: Climate Change 2022 - Impacts, Adaptation and Vulnerability, Africa Chapter
5. **Douglas et al. (2022)**: Urban flooding in Lagos, Nigeria - Nature Climate Change
6. **Mavume et al. (2021)**: Flood risk in Maputo, Mozambique - International Journal of Disaster Risk Reduction
7. **Ouma & Tateishi (2020)**: Urban flood vulnerability in Nairobi - Remote Sensing
8. **World Bank GFDRR (2023)**: Understanding Risk in African Urban Centers

## Usage

### Load with Hugging Face Datasets
```python
from datasets import load_dataset

# Load all scenarios
ds = load_dataset("electricsheepafrica/african-flood-risk-urban-mapping")

# Load specific scenario
ds_baseline = load_dataset("electricsheepafrica/african-flood-risk-urban-mapping", "historical_baseline")
ds_amplified = load_dataset("electricsheepafrica/african-flood-risk-urban-mapping", "climate_amplified")
ds_resilient = load_dataset("electricsheepafrica/african-flood-risk-urban-mapping", "resilient_infrastructure")

# Access data
df = ds["train"].to_pandas()
```

### Load from CSV
```python
import pandas as pd

# Full dataset
df = pd.read_csv("african_flood_risk_urban_mapping.csv")

# Scenario-specific
df_baseline = pd.read_csv("historical_baseline.csv")
df_amplified = pd.read_csv("climate_amplified.csv")
df_resilient = pd.read_csv("resilient_infrastructure.csv")
```

### Example Analysis
```python
# Compare flood frequency across scenarios for Lagos
lagos = df[df["city"] == "Lagos"]
lagos.groupby("scenario")["annual_flood_frequency_per_1000"].mean()

# Identify highest-risk districts
top_risk = df.nlargest(20, "composite_flood_risk_score")[["city", "district", "composite_flood_risk_score", "scenario"]]

# Economic damage by region
df.groupby(["region", "scenario"])["annual_expected_damage_usd"].sum()
```

## Reproduction

```bash
pip install -r requirements.txt
python generate_dataset.py
python validate_dataset.py
```

## Validation

The dataset passes 63 validation checks including:
- Record count >= 22,000
- All 18 cities and 3 scenarios present
- All 67 required columns present
- Value range checks for all indices and fractions
- Return period depth ordering (10yr < 25yr < 50yr < 100yr < 200yr)
- Scenario plausibility (amplified > baseline > resilient for risk metrics)
- No null values in critical columns
- Population consistency checks
- Non-negative economic damage values

## License

CC-BY-4.0

## Citation

```
@dataset{african_flood_risk_urban_mapping_2026,
title = {African Flood Risk Urban Mapping},
author = {Electric Sheep Africa},
year = {2026},
url = {https://huggingface.co/datasets/electricsheepafrica/african-flood-risk-urban-mapping},
note = {Parameters calibrated from World Bank, AfDB, UNDRR, and IPCC AR6 sources}
}
```