Upload folder using huggingface_hub

README.md

@@ -0,0 +1,107 @@
+---
+license: cc-by-4.0
+task_categories:
+  - tabular-classification
+  - tabular-regression
+language:
+  - en
+tags:
+  - climate
+  - environment
+  - africa
+  - synthetic-data
+  - sub-saharan-africa
+  - climate-vulnerability
+  - climate-adaptation
+  - climate-risk
+size_categories:
+  - 10K<n<100K
+---
+
+# Climate Vulnerability and Exposure - Africa
+
+A comprehensive synthetic dataset assessing climate vulnerability, exposure indices, and adaptive capacity across Sub-Saharan African countries.
+
+## Dataset Description
+
+This dataset provides detailed records of climate vulnerability assessments across 15 Sub-Saharan African countries. It includes exposure metrics, hazard probabilities, adaptive capacity indicators, and climate policy frameworks. The data enables analysis of climate-nutrition linkages through malnutrition burden stratification.
+
+### Key Statistics
+
+| Metric | Value |
+|--------|-------|
+| Total Records | 14,998 |
+| Countries Covered | 15 |
+| Time Period | 2018-2025 |
+| Files | 3 (high_burden, moderate_burden, low_burden) |
+| Features | 29 |
+
+### Countries Included
+DRC, Ethiopia, Ghana, Kenya, Malawi, Mali, Mozambique, Niger, Nigeria, Rwanda, Senegal, South Africa, Tanzania, Uganda, Zambia
+
+## Column Descriptions
+
+| Column | Type | Description |
+|--------|------|-------------|
+| record_id | string | Unique identifier (CLIM_XXXXXX) |
+| country | string | African country name |
+| year | int | Year of observation (2018-2025) |
+| climate_zone | string | Climate classification |
+| geographic_exposure_index | float | Geographic exposure to climate hazards (0-1) |
+| coastal_proximity | int | Binary coastal proximity (0/1) |
+| elevation_category | string | Elevation classification |
+| temperature_anomaly_c | float | Temperature deviation from baseline (°C) |
+| precipitation_variability_pct | float | Precipitation variability (%) |
+| extreme_heat_days_year | int | Annual extreme heat days |
+| heat_stress_index | float | Heat stress composite index |
+| drought_probability | float | Probability of drought (0-1) |
+| flood_probability | float | Probability of flooding (0-1) |
+| climate_hazard_index | float | Composite hazard index (0-1) |
+| gdp_per_capita_usd | float | GDP per capita (USD) |
+| agriculture_gdp_share_pct | float | Agriculture share of GDP (%) |
+| infrastructure_quality_index | float | Infrastructure quality (0-1) |
+| healthcare_access_index | float | Healthcare access (0-1) |
+| early_warning_coverage_pct | float | Early warning system coverage (%) |
+| social_protection_coverage_pct | float | Social protection coverage (%) |
+| adaptive_capacity_index | float | Overall adaptive capacity (0-1) |
+| climate_policy_strength | float | Climate policy strength (0-1) |
+| ndc_ambition_score | float | NDC ambition score (0-1) |
+| climate_finance_access_million_usd | float | Climate finance accessed (million USD) |
+| vulnerability_index | float | Composite vulnerability index (0-1) |
+| exposure_score | float | Climate exposure score (0-1) |
+| risk_category | string | Risk classification |
+| population_at_risk_millions | float | Population at climate risk (millions) |
+| potential_economic_loss_pct_gdp | float | Potential economic loss (% GDP) |
+
+### Climate Zones
+- tropical_wet, tropical_dry, semi_arid, arid, mediterranean, highland
+
+### Risk Categories
+- Low, Moderate, High, Very_High
+
+## Usage Example
+
+```python
+import pandas as pd
+
+# Load high burden dataset
+df = pd.read_csv('climate_vulnerability_exposure_africa_high_burden.csv')
+
+# Analyze vulnerability by country
+vulnerability = df.groupby('country')['vulnerability_index'].mean()
+print(vulnerability.sort_values(ascending=False))
+
+# Filter very high risk areas
+very_high_risk = df[df['risk_category'] == 'Very_High']
+
+# Correlation analysis
+correlation = df[['adaptive_capacity_index', 'vulnerability_index']].corr()
+```
+
+## Research Sources
+
+This synthetic dataset is inspired by and aligned with data from:
+- [ND-GAIN Country Index](https://gain.nd.edu/our-work/country-index/)
+- [INFORM Risk Index](https://drmkc.jrc.ec.europa.eu/inform-index)
+- [World Bank Climate Change Knowledge Portal](https://climateknowledgeportal.worldbank.org/)
+- [IPCC Climate Reports](https://www.ipcc.ch/)

generate_dataset.py

@@ -0,0 +1,225 @@
+"""
+Climate Vulnerability and Exposure Dataset Generator for Africa
+
+================================================================================
+PARAMETER EVIDENCE TABLE
+================================================================================
+| Parameter                    | Value/Range           | Source                   |
+|------------------------------|-----------------------|--------------------------|
+| Africa warming rate          | 1.5x global average   | IPCC AR6 2023            |
+| Extreme heat events increase | 4x since 1990         | WMO State of Climate     |
+| Climate-related displacement | 7M annually           | IDMC 2024                |
+| Agricultural GDP loss risk   | 10-15% by 2050        | World Bank 2023          |
+| Adaptive capacity index      | 0.25-0.45 (Africa)    | ND-GAIN 2024             |
+| Climate finance gap          | $250B annually        | AfDB 2024                |
+| Coastal flood exposure       | 54M people            | WRI Aqueduct 2024        |
+| Rainfall variability         | ±40% from mean        | ClimDev-Africa 2024      |
+================================================================================
+
+DAG Structure:
+geographic_exposure -> climate_hazard_probability -> vulnerability_index
+economic_capacity -> adaptive_capacity
+infrastructure_resilience -> exposure_score
+institutional_capacity -> response_capability
+temperature_trends -> heat_stress_index
+precipitation_patterns -> drought_flood_risk
+"""
+
+import numpy as np
+import pandas as pd
+from pathlib import Path
+
+COUNTRIES = [
+    "Kenya", "Uganda", "Nigeria", "Ghana", "Tanzania", "Ethiopia",
+    "Malawi", "Zambia", "Senegal", "Rwanda", "Niger", "Mali",
+    "DRC", "Mozambique", "South Africa"
+]
+
+YEARS = list(range(2018, 2026))
+
+COUNTRY_VULNERABILITY_BASE = {
+    "Niger": {"exposure": 0.85, "adaptive_capacity": 0.25},
+    "Mali": {"exposure": 0.82, "adaptive_capacity": 0.28},
+    "DRC": {"exposure": 0.75, "adaptive_capacity": 0.30},
+    "Mozambique": {"exposure": 0.78, "adaptive_capacity": 0.32},
+    "Ethiopia": {"exposure": 0.72, "adaptive_capacity": 0.35},
+    "Malawi": {"exposure": 0.70, "adaptive_capacity": 0.33},
+    "Tanzania": {"exposure": 0.65, "adaptive_capacity": 0.38},
+    "Uganda": {"exposure": 0.62, "adaptive_capacity": 0.40},
+    "Kenya": {"exposure": 0.58, "adaptive_capacity": 0.42},
+    "Zambia": {"exposure": 0.60, "adaptive_capacity": 0.40},
+    "Ghana": {"exposure": 0.50, "adaptive_capacity": 0.48},
+    "Senegal": {"exposure": 0.55, "adaptive_capacity": 0.45},
+    "Nigeria": {"exposure": 0.52, "adaptive_capacity": 0.44},
+    "Rwanda": {"exposure": 0.48, "adaptive_capacity": 0.52},
+    "South Africa": {"exposure": 0.42, "adaptive_capacity": 0.58},
+}
+
+CLIMATE_ZONES = ["arid", "semi_arid", "tropical_wet", "tropical_dry", "mediterranean", "highland"]
+
+def generate_dataset(scenario: str, n_samples: int, seed: int) -> pd.DataFrame:
+    rng = np.random.default_rng(seed)
+    
+    burden_mod = {"low_burden": 0.7, "moderate_burden": 1.0, "high_burden": 1.4}[scenario]
+    
+    countries = rng.choice(COUNTRIES, size=n_samples)
+    years = rng.choice(YEARS, size=n_samples)
+    
+    data = {
+        "record_id": [f"CLIM_{i:06d}" for i in range(n_samples)],
+        "country": countries,
+        "year": years,
+        "climate_zone": rng.choice(CLIMATE_ZONES, size=n_samples, 
+                                   p=[0.15, 0.25, 0.20, 0.20, 0.05, 0.15]),
+    }
+    
+    zone = data["climate_zone"]
+    base_exposure = np.array([COUNTRY_VULNERABILITY_BASE[c]["exposure"] for c in countries])
+    base_adaptive = np.array([COUNTRY_VULNERABILITY_BASE[c]["adaptive_capacity"] for c in countries])
+    
+    year_factor = (years - 2018) * 0.02 * burden_mod
+    data["geographic_exposure_index"] = np.clip(
+        base_exposure * (1 + year_factor) + rng.normal(0, 0.05, n_samples), 0.1, 1.0
+    ).round(3)
+    
+    data["coastal_proximity"] = rng.choice([0, 1], size=n_samples, p=[0.6, 0.4])
+    data["elevation_category"] = rng.choice(["lowland", "mid_elevation", "highland"], 
+                                            size=n_samples, p=[0.4, 0.35, 0.25])
+    
+    zone_temp_factor = {"arid": 1.2, "semi_arid": 1.1, "tropical_wet": 0.9, 
+                        "tropical_dry": 1.0, "mediterranean": 0.8, "highland": 0.7}
+    temp_factors = np.array([zone_temp_factor[z] for z in zone])
+    
+    data["temperature_anomaly_c"] = np.clip(
+        (0.8 + year_factor) * temp_factors * rng.uniform(0.5, 1.5, n_samples), 0.2, 3.5
+    ).round(2)
+    
+    zone_precip_factor = {"arid": 1.3, "semi_arid": 1.2, "tropical_wet": 1.1, 
+                          "tropical_dry": 1.4, "mediterranean": 1.2, "highland": 1.0}
+    precip_factors = np.array([zone_precip_factor[z] for z in zone])
+    
+    data["precipitation_variability_pct"] = np.clip(
+        25 + year_factor * 10 * precip_factors + rng.normal(0, 5, n_samples), 10, 60
+    ).round(1)
+    
+    data["extreme_heat_days_year"] = np.clip(
+        (20 + data["temperature_anomaly_c"] * 15) * burden_mod * 
+        rng.uniform(0.8, 1.2, n_samples), 5, 150
+    ).astype(int)
+    
+    data["heat_stress_index"] = np.clip(
+        (data["temperature_anomaly_c"] / 2 + data["extreme_heat_days_year"] / 50) * 
+        burden_mod, 0, 10
+    ).round(2)
+    
+    data["drought_probability"] = np.clip(
+        data["geographic_exposure_index"] * 0.5 + 
+        data["precipitation_variability_pct"] / 200 + 
+        rng.uniform(-0.1, 0.1, n_samples), 0, 1
+    ).round(3)
+    
+    data["flood_probability"] = np.clip(
+        (data["coastal_proximity"] * 0.2 + 
+         data["precipitation_variability_pct"] / 150) * 
+        burden_mod + rng.uniform(0, 0.15, n_samples), 0, 0.8
+    ).round(3)
+    
+    data["climate_hazard_index"] = np.clip(
+        (data["drought_probability"] * 0.4 + 
+         data["flood_probability"] * 0.3 + 
+         data["heat_stress_index"] / 20) * burden_mod, 0, 1
+    ).round(3)
+    
+    data["gdp_per_capita_usd"] = np.clip(
+        rng.exponential(3000, n_samples) * (1 - data["geographic_exposure_index"] * 0.5), 
+        300, 15000
+    ).astype(int)
+    
+    data["agriculture_gdp_share_pct"] = np.clip(
+        30 - data["gdp_per_capita_usd"] / 500 + rng.normal(0, 5, n_samples), 5, 50
+    ).round(1)
+    
+    data["infrastructure_quality_index"] = np.clip(
+        base_adaptive * 0.8 + rng.normal(0, 0.1, n_samples), 0.1, 0.9
+    ).round(3)
+    
+    data["healthcare_access_index"] = np.clip(
+        base_adaptive * 0.9 + data["gdp_per_capita_usd"] / 20000 + 
+        rng.normal(0, 0.1, n_samples), 0.1, 0.95
+    ).round(3)
+    
+    data["early_warning_coverage_pct"] = np.clip(
+        base_adaptive * 60 + rng.normal(10, 15, n_samples), 5, 90
+    ).round(1)
+    
+    data["social_protection_coverage_pct"] = np.clip(
+        data["gdp_per_capita_usd"] / 150 + rng.normal(5, 10, n_samples), 2, 60
+    ).round(1)
+    
+    data["adaptive_capacity_index"] = np.clip(
+        (data["infrastructure_quality_index"] * 0.3 + 
+         data["healthcare_access_index"] * 0.25 + 
+         data["early_warning_coverage_pct"] / 200 + 
+         data["social_protection_coverage_pct"] / 150) + 
+        rng.normal(0, 0.05, n_samples), 0.1, 0.95
+    ).round(3)
+    
+    data["climate_policy_strength"] = np.clip(
+        base_adaptive * 1.2 + rng.normal(0, 0.15, n_samples), 0.1, 0.9
+    ).round(3)
+    
+    data["ndc_ambition_score"] = np.clip(
+        data["climate_policy_strength"] * 0.8 + rng.uniform(0, 0.3, n_samples), 0.1, 1.0
+    ).round(3)
+    
+    data["climate_finance_access_million_usd"] = np.clip(
+        data["climate_policy_strength"] * 500 + rng.exponential(100, n_samples), 10, 2000
+    ).astype(int)
+    
+    data["vulnerability_index"] = np.clip(
+        (data["climate_hazard_index"] * 0.5 + 
+         data["geographic_exposure_index"] * 0.3 + 
+         (1 - data["adaptive_capacity_index"]) * 0.2) * burden_mod, 0, 1
+    ).round(3)
+    
+    data["exposure_score"] = np.clip(
+        (data["geographic_exposure_index"] + data["climate_hazard_index"]) / 2, 0, 1
+    ).round(3)
+    
+    data["risk_category"] = np.select(
+        [data["vulnerability_index"] < 0.3, 
+         data["vulnerability_index"] < 0.5, 
+         data["vulnerability_index"] < 0.7, 
+         data["vulnerability_index"] < 0.85, 
+         data["vulnerability_index"] >= 0.85],
+        ["Very_Low", "Low", "Moderate", "High", "Very_High"]
+    )
+    
+    data["population_at_risk_millions"] = np.clip(
+        data["vulnerability_index"] * 15 * burden_mod + 
+        rng.exponential(2, n_samples), 0.5, 50
+    ).round(2)
+    
+    data["potential_economic_loss_pct_gdp"] = np.clip(
+        data["vulnerability_index"] * 20 + data["agriculture_gdp_share_pct"] / 5, 1, 25
+    ).round(1)
+    
+    return pd.DataFrame(data)
+
+def main():
+    output_dir = Path(__file__).parent
+    
+    scenarios = [
+        ("low_burden", 4000, 42),
+        ("moderate_burden", 5000, 43),
+        ("high_burden", 6000, 44),
+    ]
+    
+    for scenario, n, seed in scenarios:
+        df = generate_dataset(scenario, n, seed)
+        output_file = output_dir / f"climate_vulnerability_exposure_africa_{scenario}.csv"
+        df.to_csv(output_file, index=False)
+        print(f"Generated {output_file}: {len(df)} records")
+
+if __name__ == "__main__":
+    main()