generate_dataset.py

"""Generate synthetic e-waste recycling & occupational health dataset for SSA.

Research-based parameterization:
- Agbogbloshie, Ghana = world's largest informal e-waste recycling site;
  Ghana imports ~215,000 tonnes secondhand electronics/yr (Pure Earth).
- E-waste manually dismantled by burning plastic/insulation to extract Cu,
  Au; releases Pb, Cd, Cr, Hg, PAHs, dioxins, BFRs (PMC10815197).
- Biomonitoring: 75 Agbogbloshie workers vs 40 controls showed elevated
  blood Pb, Cd, Cr, Ni, Hg (PubMed 27858271; PMC8287752).
- Children near e-waste sites: elevated BLL, DNA damage (PMC8392572).
- Respiratory symptoms & reduced lung function in burning workers
  (PMC7084368).
- No PPE or environmental protection in informal recycling
  (ScienceDirect 2024).
- Most SSA countries lack e-waste legislation; enforcement weak where
  laws exist (ScienceDirect 2025 review).
- E-waste ranked Top 10 toxic threats globally (Pure Earth 2013).
"""

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 = {
    # Agbogbloshie-type mega-site
    "mega_site_west_africa": {
        "setting_probs": {"urban_dumpsite": 0.50, "peri_urban": 0.30, "urban_market": 0.20},
        "role_probs": {"burner": 0.25, "dismantler": 0.25, "collector": 0.15,
                       "community_resident": 0.20, "child_worker": 0.15},
        # Blood Pb elevated: workers ~15-25 µg/dL (Agbogbloshie studies)
        "blood_pb_gm_worker": 18.0, "blood_pb_gsd": 2.0,
        "blood_pb_gm_community": 6.5,
        "blood_cd_gm_worker": 1.5,  # µg/L
        "blood_cd_gm_community": 0.5,
        "urine_cr_gm_worker": 3.0,  # µg/L
        "ppe_use_pct": 0.05,
        "respiratory_symptoms_burner": 0.55,
        "skin_symptoms_pct": 0.30,
        "ewaste_volume_tonnes_yr": 215000,
    },
    # Medium informal site (Lagos, Nairobi type)
    "medium_site_urban": {
        "setting_probs": {"urban_market": 0.45, "peri_urban": 0.35, "urban_dumpsite": 0.20},
        "role_probs": {"burner": 0.20, "dismantler": 0.30, "collector": 0.15,
                       "community_resident": 0.25, "child_worker": 0.10},
        "blood_pb_gm_worker": 12.0, "blood_pb_gsd": 1.9,
        "blood_pb_gm_community": 5.0,
        "blood_cd_gm_worker": 1.0,
        "blood_cd_gm_community": 0.4,
        "urine_cr_gm_worker": 2.0,
        "ppe_use_pct": 0.10,
        "respiratory_symptoms_burner": 0.45,
        "skin_symptoms_pct": 0.22,
        "ewaste_volume_tonnes_yr": 50000,
    },
    # Small dispersed sites (smaller cities)
    "small_dispersed_sites": {
        "setting_probs": {"peri_urban": 0.45, "urban_market": 0.30, "rural": 0.25},
        "role_probs": {"burner": 0.15, "dismantler": 0.30, "collector": 0.20,
                       "community_resident": 0.25, "child_worker": 0.10},
        "blood_pb_gm_worker": 8.0, "blood_pb_gsd": 1.8,
        "blood_pb_gm_community": 4.0,
        "blood_cd_gm_worker": 0.7,
        "blood_cd_gm_community": 0.3,
        "urine_cr_gm_worker": 1.5,
        "ppe_use_pct": 0.08,
        "respiratory_symptoms_burner": 0.35,
        "skin_symptoms_pct": 0.18,
        "ewaste_volume_tonnes_yr": 10000,
    },
}

SCENARIO_FILES = {
    "mega_site_west_africa": "ewaste_mega_site.csv",
    "medium_site_urban": "ewaste_medium_urban.csv",
    "small_dispersed_sites": "ewaste_small_dispersed.csv",
}

EWASTE_TYPES = {"computers_monitors": 0.25, "mobile_phones": 0.20, "cables_wiring": 0.20,
                "appliances": 0.15, "batteries": 0.10, "mixed_other": 0.10}

PROCESSING_METHODS = {"open_burning": 0.35, "manual_dismantling": 0.30, "acid_bath": 0.10,
                      "hammer_chisel": 0.15, "sorting_only": 0.10}


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(28, 12), 8, 60))
        sex = rng.choice(["male", "female"], p=[0.72, 0.28])
        role = _choice(rng, params["role_probs"])
        is_worker = int(role in ("burner", "dismantler", "collector"))
        is_child = int(age < 18)
        years_exposure = int(np.clip(rng.normal(5, 3), 0, 25)) if is_worker else int(np.clip(rng.normal(3, 2), 0, 15))

        ewaste_type = _choice(rng, EWASTE_TYPES)
        processing_method = _choice(rng, PROCESSING_METHODS) if is_worker else "none"
        is_burner = int(role == "burner" or processing_method == "open_burning")

        ppe_use = int(is_worker and rng.random() < params["ppe_use_pct"])
        gloves = int(ppe_use and rng.random() < 0.60)
        mask = int(ppe_use and rng.random() < 0.30)
        hours_per_day = float(np.clip(rng.normal(8, 2), 2, 14)) if is_worker else 0

        # Biomarkers (PubMed 27858271; PMC8287752)
        if is_worker:
            blood_pb = float(np.clip(
                rng.lognormal(np.log(params["blood_pb_gm_worker"]), np.log(params["blood_pb_gsd"])),
                1, 100,
            ))
            blood_cd = float(np.clip(
                rng.lognormal(np.log(params["blood_cd_gm_worker"]), 0.7), 0.1, 20,
            ))
            urine_cr = float(np.clip(
                rng.lognormal(np.log(params["urine_cr_gm_worker"]), 0.6), 0.1, 30,
            ))
        else:
            blood_pb = float(np.clip(
                rng.lognormal(np.log(params["blood_pb_gm_community"]), np.log(1.7)),
                0.5, 50,
            ))
            blood_cd = float(np.clip(
                rng.lognormal(np.log(params["blood_cd_gm_community"]), 0.6), 0.05, 10,
            ))
            urine_cr = float(np.clip(
                rng.lognormal(np.log(max(params["urine_cr_gm_worker"] * 0.3, 0.3)), 0.5), 0.05, 15,
            ))

        if ppe_use:
            blood_pb *= 0.8
            blood_cd *= 0.85

        elevated_pb = int(blood_pb >= 10)
        elevated_cd = int(blood_cd >= 1.0)

        # Health outcomes
        # Respiratory (PMC7084368: burning workers high risk)
        resp_risk = params["respiratory_symptoms_burner"] if is_burner else 0.10
        cough_chronic = int(rng.random() < resp_risk * 0.8)
        wheeze = int(rng.random() < resp_risk * 0.5)
        dyspnoea = int(rng.random() < resp_risk * 0.4)
        reduced_fev1 = int(is_burner and rng.random() < 0.25)
        any_respiratory = int(cough_chronic or wheeze or dyspnoea or reduced_fev1)

        # Skin/dermal (PMC10815197)
        skin_rash = int(rng.random() < params["skin_symptoms_pct"])
        burns_injury = int(is_burner and rng.random() < 0.15)
        eye_irritation = int(is_burner and rng.random() < 0.30)

        # Neurological (lead-related)
        headache = int(rng.random() < np.clip(0.10 + blood_pb * 0.005, 0, 0.5))
        fatigue = int(rng.random() < np.clip(0.15 + blood_pb * 0.004, 0, 0.5))
        child_developmental = int(is_child and blood_pb > 5 and rng.random() < 0.20)

        # Kidney (Cd-related)
        proteinuria = int(blood_cd > 1.0 and rng.random() < 0.15)

        # DNA damage (PMC8392572)
        dna_damage_risk = int(rng.random() < np.clip(0.05 + years_exposure * 0.01 + is_burner * 0.10, 0, 0.4))

        # Legislation & formalization
        ewaste_legislation_exists = int(rng.random() < 0.25)
        formal_recycler = int(is_worker and rng.random() < 0.05)
        health_screening_access = int(rng.random() < 0.08)

        record = {
            "record_id": f"{name[:3].upper()}-{idx:05d}",
            "scenario": name,
            "year": year,
            "setting": setting,
            "age": age,
            "sex": sex,
            "role": role,
            "is_worker": is_worker,
            "is_child": is_child,
            "years_exposure": years_exposure,
            "ewaste_type": ewaste_type,
            "processing_method": processing_method,
            "is_burner": is_burner,
            "ppe_use": ppe_use,
            "gloves": gloves,
            "mask": mask,
            "hours_per_day": round(hours_per_day, 1),
            "blood_pb_ugdL": round(blood_pb, 1),
            "blood_cd_ugL": round(blood_cd, 2),
            "urine_cr_ugL": round(urine_cr, 2),
            "elevated_pb": elevated_pb,
            "elevated_cd": elevated_cd,
            "cough_chronic": cough_chronic,
            "wheeze": wheeze,
            "dyspnoea": dyspnoea,
            "reduced_fev1": reduced_fev1,
            "any_respiratory": any_respiratory,
            "skin_rash": skin_rash,
            "burns_injury": burns_injury,
            "eye_irritation": eye_irritation,
            "headache": headache,
            "fatigue": fatigue,
            "child_developmental": child_developmental,
            "proteinuria": proteinuria,
            "dna_damage_risk": dna_damage_risk,
            "ewaste_legislation": ewaste_legislation_exists,
            "formal_recycler": formal_recycler,
            "health_screening_access": health_screening_access,
        }
        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()