electricsheepafrica/africa-disaster-risk-all

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

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

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

# 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

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}
}