Upload folder using huggingface_hub

README.md

@@ -0,0 +1,88 @@
+---
+license: cc-by-4.0
+task_categories:
+  - tabular-classification
+language:
+  - en
+tags:
+  - environmental-health
+  - radon
+  - indoor-radiation
+  - lung-cancer
+  - building-materials
+  - synthetic
+  - sub-saharan-africa
+pretty_name: Radon & Indoor Radiation Exposure (SSA)
+size_categories:
+  - 10K<n<100K
+configs:
+  - config_name: granite_geology_rural
+    data_files: data/radon_granite_rural.csv
+    default: true
+  - config_name: urban_residential
+    data_files: data/radon_urban_residential.csv
+  - config_name: occupational_underground
+    data_files: data/radon_occupational.csv
+---
+
+# Radon & Indoor Radiation Exposure in Sub-Saharan Africa
+
+## Abstract
+
+Synthetic dataset modelling indoor radon concentrations, building characteristics, and lung cancer risk across three settings in SSA. Radon is the second leading cause of lung cancer globally; WHO recommends 100 Bq/m³ reference level. Granite geology, poor ventilation, and ground-level dwellings increase concentrations.
+
+### Scenarios
+
+- **Granite Geology Rural**: Communities on granite bedrock with mean radon ~120 Bq/m³.
+- **Urban Residential**: Mixed urban buildings with mean ~60 Bq/m³.
+- **Occupational Underground**: Mines, tunnels, basements with mean ~250 Bq/m³.
+
+## Parameterization Evidence
+
+| Parameter | Value | Source | Year |
+| --- | --- | --- | --- |
+| Radon causes 3-14% of lung cancers; 100 Bq/m³ ref | Guideline | WHO Fact Sheet | 2023 |
+| Indoor radon in Africa: critical review | SSA data | PMC12277776 | 2025 |
+| ~50% of human radiation exposure from radon | Burden | PMC12081354 | 2025 |
+| Nigerian buildings radon monitoring | SSA data | PubMed 40334468 | 2025 |
+| SA community near granite: alpha particle damage | SSA data | PMC12331818 | 2025 |
+
+## Validation
+
+![Validation Report](validation_report.png)
+
+## Usage
+
+```python
+from datasets import load_dataset
+ds = load_dataset("electricsheepafrica/radon-indoor-radiation", "granite_geology_rural")
+```
+
+## Limitations
+
+- Synthetic data; not for clinical decision-making.
+- Radon concentrations vary seasonally and diurnally; dataset captures annual average.
+- Limited real measurement data from SSA to validate distributions.
+
+## References
+
+1. WHO. Radon and Health Fact Sheet. 2023.
+2. PMC12277776. Indoor radon exposure in Africa: critical review. 2025.
+3. PMC12081354. Lung cancer attributed to residential radon. 2025.
+4. PMC12331818. Indoor radon in SA community near granite. 2025.
+
+## Citation
+
+```bibtex
+@dataset{electricsheepafrica_radon_indoor_radiation_2025,
+  title={Radon and Indoor Radiation Exposure in Sub-Saharan Africa},
+  author={Electric Sheep Africa},
+  year={2025},
+  publisher={HuggingFace},
+  url={https://huggingface.co/datasets/electricsheepafrica/radon-indoor-radiation}
+}
+```
+
+## License
+
+CC-BY-4.0

generate_dataset.py

@@ -0,0 +1,185 @@
+"""Generate synthetic radon & indoor radiation exposure dataset for SSA.
+
+Research-based parameterization:
+- WHO Fact Sheet: Radon causes 3-14% of lung cancers; reference level
+  100 Bq/m³, action level 300 Bq/m³.
+- PMC12277776: Indoor radon exposure in Africa critical review; growing
+  concern; classified IARC Group 1 carcinogen.
+- PMC12081354: Radon = ~50% of human radiation exposure; originates from
+  granite, brick, sand, cement, gypsum.
+- PubMed 40334468: Nigerian buildings (homes, schools, workplaces)
+  monitored; significant public health concern.
+- PMC12331818: SA community near granite geology; alpha particles from
+  radon daughters damage lung cells.
+- Second leading cause of lung cancer globally after smoking.
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+
+SEED = 42
+N_PER_SCENARIO = 10_000
+
+YEAR_RANGE = np.arange(2010, 2025)
+YEAR_WEIGHTS = np.linspace(0.85, 1.3, len(YEAR_RANGE))
+YEAR_WEIGHTS = YEAR_WEIGHTS / YEAR_WEIGHTS.sum()
+
+SCENARIOS = {
+    "granite_geology_rural": {
+        "setting_probs": {"rural_granite": 0.50, "peri_urban": 0.30, "mining_town": 0.20},
+        "building_probs": {"mud_brick": 0.30, "concrete_block": 0.25,
+                           "stone_granite": 0.25, "corrugated_iron": 0.15, "modern": 0.05},
+        "radon_gm": 120, "radon_gsd": 2.2,
+        "ventilation_poor_pct": 0.55,
+        "smoking_prev": 0.15,
+        "lung_cancer_base": 0.003,
+        "measurement_pct": 0.02,
+    },
+    "urban_residential": {
+        "setting_probs": {"urban_formal": 0.40, "urban_informal": 0.35, "peri_urban": 0.25},
+        "building_probs": {"concrete_block": 0.40, "brick": 0.25,
+                           "corrugated_iron": 0.15, "modern": 0.15, "mud_brick": 0.05},
+        "radon_gm": 60, "radon_gsd": 2.0,
+        "ventilation_poor_pct": 0.35,
+        "smoking_prev": 0.20,
+        "lung_cancer_base": 0.002,
+        "measurement_pct": 0.05,
+    },
+    "occupational_underground": {
+        "setting_probs": {"underground_mine": 0.45, "tunnel_cave": 0.20,
+                          "basement_building": 0.20, "industrial": 0.15},
+        "building_probs": {"underground": 0.45, "concrete_block": 0.25,
+                           "stone_granite": 0.15, "modern": 0.10, "other": 0.05},
+        "radon_gm": 250, "radon_gsd": 2.5,
+        "ventilation_poor_pct": 0.45,
+        "smoking_prev": 0.20,
+        "lung_cancer_base": 0.005,
+        "measurement_pct": 0.08,
+    },
+}
+
+SCENARIO_FILES = {
+    "granite_geology_rural": "radon_granite_rural.csv",
+    "urban_residential": "radon_urban_residential.csv",
+    "occupational_underground": "radon_occupational.csv",
+}
+
+
+def _choice(rng, prob_map):
+    keys = list(prob_map.keys())
+    weights = np.array(list(prob_map.values()), dtype=float)
+    weights = weights / weights.sum()
+    return rng.choice(keys, p=weights)
+
+
+def _simulate_scenario(name, params, seed):
+    rng = np.random.default_rng(seed)
+    records = []
+
+    for idx in range(N_PER_SCENARIO):
+        year = int(rng.choice(YEAR_RANGE, p=YEAR_WEIGHTS))
+        setting = _choice(rng, params["setting_probs"])
+        age = int(np.clip(rng.normal(40, 15), 18, 80))
+        sex = rng.choice(["male", "female"], p=[0.50, 0.50])
+        building_type = _choice(rng, params["building_probs"])
+
+        # Ventilation (poor ventilation increases radon accumulation)
+        ventilation_poor = int(rng.random() < params["ventilation_poor_pct"])
+        floor_level = rng.choice(["ground_basement", "first", "upper"],
+                                 p=[0.60, 0.25, 0.15])
+
+        # Radon concentration (Bq/m³)
+        vent_mult = 1.5 if ventilation_poor else 1.0
+        floor_mult = 1.3 if floor_level == "ground_basement" else 0.8 if floor_level == "upper" else 1.0
+        geology_mult = 1.3 if building_type in ("stone_granite", "underground") else 1.0
+
+        radon_bqm3 = float(np.clip(
+            rng.lognormal(np.log(params["radon_gm"] * vent_mult * floor_mult * geology_mult),
+                          np.log(params["radon_gsd"])),
+            5, 3000))
+
+        above_who_100 = int(radon_bqm3 > 100)
+        above_action_300 = int(radon_bqm3 > 300)
+
+        # Exposure
+        hours_indoors_day = float(np.clip(rng.normal(16, 3), 8, 23))
+        exposure_years = int(np.clip(rng.normal(15, 8), 1, 50))
+        occupancy_factor = hours_indoors_day / 24
+        cumulative_exposure = float(radon_bqm3 * occupancy_factor * exposure_years)
+
+        smoking = int(rng.random() < params["smoking_prev"])
+        # Smoking + radon synergy: multiplicative risk (WHO)
+        risk_mult = (cumulative_exposure / 500) * (5.0 if smoking else 1.0)
+
+        # Lung cancer (WHO: radon = second leading cause)
+        lung_cancer = int(age >= 40 and rng.random() < np.clip(
+            params["lung_cancer_base"] * risk_mult, 0, 0.05))
+
+        # Respiratory symptoms
+        chronic_cough = int(rng.random() < np.clip(0.05 + risk_mult * 0.02, 0, 0.20))
+        dyspnoea = int(rng.random() < np.clip(0.04 + risk_mult * 0.01, 0, 0.15))
+
+        # Measurement & mitigation
+        radon_measured = int(rng.random() < params["measurement_pct"])
+        aware_of_radon = int(rng.random() < 0.05)
+        mitigation_installed = int(above_action_300 and radon_measured and rng.random() < 0.10)
+        ventilation_improved = int(radon_measured and rng.random() < 0.15)
+        sub_slab_depressurization = int(mitigation_installed and rng.random() < 0.20)
+
+        # Building characteristics
+        sealed_floor = int(rng.random() < 0.30)
+        cracks_in_floor = int(rng.random() < 0.40)
+        uranium_geology = int(setting in ("underground_mine", "rural_granite", "mining_town") and
+                              rng.random() < 0.30)
+
+        any_health_effect = int(lung_cancer or chronic_cough or dyspnoea)
+
+        record = {
+            "record_id": f"{name[:3].upper()}-{idx:05d}",
+            "scenario": name,
+            "year": year,
+            "setting": setting,
+            "age": age,
+            "sex": sex,
+            "building_type": building_type,
+            "floor_level": floor_level,
+            "ventilation_poor": ventilation_poor,
+            "radon_bqm3": round(radon_bqm3, 1),
+            "above_who_100": above_who_100,
+            "above_action_300": above_action_300,
+            "hours_indoors_day": round(hours_indoors_day, 1),
+            "exposure_years": exposure_years,
+            "cumulative_exposure": round(cumulative_exposure, 0),
+            "smoking": smoking,
+            "lung_cancer": lung_cancer,
+            "chronic_cough": chronic_cough,
+            "dyspnoea": dyspnoea,
+            "radon_measured": radon_measured,
+            "aware_of_radon": aware_of_radon,
+            "mitigation_installed": mitigation_installed,
+            "ventilation_improved": ventilation_improved,
+            "sealed_floor": sealed_floor,
+            "cracks_in_floor": cracks_in_floor,
+            "uranium_geology": uranium_geology,
+            "any_health_effect": any_health_effect,
+        }
+        records.append(record)
+
+    return pd.DataFrame(records)
+
+
+def main():
+    output_dir = Path("data")
+    output_dir.mkdir(parents=True, exist_ok=True)
+    for idx, (name, params) in enumerate(SCENARIOS.items()):
+        df = _simulate_scenario(name, params, SEED + idx * 211)
+        df.to_csv(output_dir / SCENARIO_FILES[name], index=False)
+        print(f"Saved {name} -> {SCENARIO_FILES[name]}")
+
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -0,0 +1,3 @@
+numpy>=1.24
+pandas>=2.0
+matplotlib>=3.7

validate_dataset.py

@@ -0,0 +1,92 @@
+"""Validate synthetic radon & indoor radiation exposure dataset."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import matplotlib.pyplot as plt
+import pandas as pd
+
+SCENARIO_FILES = {
+    "granite_geology_rural": "radon_granite_rural.csv",
+    "urban_residential": "radon_urban_residential.csv",
+    "occupational_underground": "radon_occupational.csv",
+}
+
+COLORS = {"granite_geology_rural": "#e6550d", "urban_residential": "#756bb1", "occupational_underground": "#31a354"}
+
+
+def load_data() -> pd.DataFrame:
+    frames = []
+    for scenario, filename in SCENARIO_FILES.items():
+        df = pd.read_csv(Path("data") / filename)
+        frames.append(df)
+    return pd.concat(frames, ignore_index=True)
+
+
+def plot_validation(df: pd.DataFrame, output_path: Path) -> None:
+    fig, axes = plt.subplots(4, 2, figsize=(14, 16))
+    axes = axes.flatten()
+
+    for s in SCENARIO_FILES:
+        subset = df[df["scenario"] == s]
+        axes[0].hist(subset["radon_bqm3"], bins=40, alpha=0.5, color=COLORS[s], label=s, range=(0, 800))
+    axes[0].axvline(100, color="red", ls="--", lw=1, label="WHO 100 Bq/m³")
+    axes[0].axvline(300, color="orange", ls="--", lw=1, label="Action 300 Bq/m³")
+    axes[0].set_title("Indoor Radon Distribution (Bq/m³)")
+    axes[0].legend(fontsize=6)
+
+    exc_cols = ["above_who_100", "above_action_300"]
+    exc = df.groupby("scenario")[exc_cols].mean() * 100
+    exc.plot(kind="bar", ax=axes[1])
+    axes[1].set_title("WHO & Action Level Exceedance (%)")
+    axes[1].legend(fontsize=7)
+
+    health_cols = ["lung_cancer", "chronic_cough", "dyspnoea"]
+    health = df.groupby("scenario")[health_cols].mean() * 100
+    health.plot(kind="bar", ax=axes[2])
+    axes[2].set_title("Health Outcomes (%)")
+    axes[2].legend(fontsize=7)
+
+    for s in SCENARIO_FILES:
+        subset = df[df["scenario"] == s]
+        axes[3].scatter(subset["radon_bqm3"], subset["lung_cancer"],
+                        s=4, alpha=0.05, color=COLORS[s], label=s)
+    axes[3].set_title("Radon vs Lung Cancer")
+    axes[3].legend(fontsize=7)
+
+    bld = df.groupby(["scenario", "building_type"]).size().groupby(level=0).apply(lambda s: s / s.sum())
+    bld.unstack().plot(kind="bar", stacked=True, ax=axes[4])
+    axes[4].set_title("Building Type Distribution")
+    axes[4].legend(fontsize=5)
+
+    flr = df.groupby(["scenario", "floor_level"]).size().groupby(level=0).apply(lambda s: s / s.sum())
+    flr.unstack().plot(kind="bar", stacked=True, ax=axes[5])
+    axes[5].set_title("Floor Level Distribution")
+    axes[5].legend(fontsize=7)
+
+    risk_cols = ["ventilation_poor", "cracks_in_floor", "smoking", "uranium_geology"]
+    risk = df.groupby("scenario")[risk_cols].mean() * 100
+    risk.plot(kind="bar", ax=axes[6])
+    axes[6].set_title("Risk Factors (%)")
+    axes[6].legend(fontsize=6)
+
+    mit_cols = ["radon_measured", "aware_of_radon", "mitigation_installed", "ventilation_improved"]
+    mit = df.groupby("scenario")[mit_cols].mean() * 100
+    mit.plot(kind="bar", ax=axes[7])
+    axes[7].set_title("Measurement & Mitigation (%)")
+    axes[7].legend(fontsize=6)
+
+    plt.tight_layout()
+    fig.savefig(output_path, dpi=200)
+    plt.close(fig)
+
+
+def main() -> None:
+    df = load_data()
+    plot_validation(df, Path("validation_report.png"))
+    print("Saved validation_report.png")
+
+
+if __name__ == "__main__":
+    main()