Upload folder using huggingface_hub

README.md

@@ -0,0 +1,75 @@
+---
+license: cc-by-4.0
+task_categories:
+  - tabular-classification
+language:
+  - en
+tags:
+  - substandard-falsified-medicines
+  - antibiotics
+  - AMR
+  - drug-quality
+  - synthetic
+  - sub-saharan-africa
+pretty_name: Antibiotic Quality & AMR Acceleration (SSA)
+size_categories:
+  - 10K<n<100K
+configs:
+  - config_name: community_otc_access
+    data_files: data/antibiotic_community.csv
+    default: true
+  - config_name: hospital_referral
+    data_files: data/antibiotic_hospital.csv
+  - config_name: cross_border_unregulated
+    data_files: data/antibiotic_cross_border.csv
+---
+
+# Antibiotic Quality & AMR Acceleration in Sub-Saharan Africa
+
+## Abstract
+
+Synthetic dataset modelling antibiotic quality, substandard/falsified prevalence, antibiotic use patterns, and AMR outcomes across three market settings in SSA. SF antibiotics are neglected drivers of AMR; sulfamethoxazole-trimethoprim, ampicillin, amoxicillin, ciprofloxacin have highest failure rates in Africa.
+
+## Parameterization Evidence
+
+| Parameter | Value | Source | Year |
+| --- | --- | --- | --- |
+| SF antibiotics = neglected drivers of AMR | Association | BMJ Global Health | 2022 |
+| Cotrimoxazole highest failure; amoxicillin, cipro follow | Prevalence | CIDRAP / BMJ GH | 2022 |
+| SF antibiotics quantitatively associated with AMR | Mechanism | BMJ Global Health | 2025 |
+| SF antimicrobials in East Africa; evolving problem | SSA data | PLOS ONE | 2024 |
+| API issues most frequent quality problem | Quality | CIDRAP | 2022 |
+
+## Validation
+
+![Validation Report](validation_report.png)
+
+## Usage
+
+```python
+from datasets import load_dataset
+ds = load_dataset("electricsheepafrica/antibiotic-quality-amr", "community_otc_access")
+```
+
+## References
+
+1. BMJ Global Health. SF antibiotics: neglected drivers of AMR. 2022.
+2. BMJ Global Health. SF antibiotics associated with AMR prevalence. 2025.
+3. PLOS ONE. SF antimicrobials in East Africa. 2024.
+4. Frontiers. Activities to reduce SF antibiotics in Africa. 2025.
+
+## Citation
+
+```bibtex
+@dataset{electricsheepafrica_antibiotic_quality_amr_2025,
+  title={Antibiotic Quality and AMR Acceleration in Sub-Saharan Africa},
+  author={Electric Sheep Africa},
+  year={2025},
+  publisher={HuggingFace},
+  url={https://huggingface.co/datasets/electricsheepafrica/antibiotic-quality-amr}
+}
+```
+
+## License
+
+CC-BY-4.0

generate_dataset.py

@@ -0,0 +1,202 @@
+"""Generate synthetic antibiotic quality & AMR acceleration dataset for SSA.
+
+Research-based parameterization:
+- BMJ Global Health (2022): SF antibiotics are neglected drivers of AMR;
+  sulfamethoxazole-trimethoprim highest failure frequency, followed by
+  ampicillin, amoxicillin, ciprofloxacin, tetracycline.
+- BMJ Global Health (2025): SF antibiotics quantitatively associated
+  with AMR prevalence.
+- PLOS ONE (2024): SF antimicrobials in East Africa; dynamic and
+  evolving problem.
+- Frontiers (2025): SF antibiotics particularly prevalent in Africa;
+  contributes to AMR burden.
+- CIDRAP: API issues most frequent quality problem; highest failure
+  rates in Africa and Asia.
+"""
+
+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 = {
+    "community_otc_access": {
+        "setting_probs": {"community_pharmacy": 0.30, "drug_shop": 0.30,
+                          "informal_vendor": 0.25, "market_stall": 0.15},
+        "antibiotic_probs": {"amoxicillin": 0.25, "cotrimoxazole": 0.20, "ciprofloxacin": 0.15,
+                             "metronidazole": 0.12, "doxycycline": 0.10, "ampicillin": 0.08,
+                             "tetracycline": 0.05, "azithromycin": 0.05},
+        "sf_prevalence": 0.30,
+        "no_prescription_pct": 0.70,
+        "incomplete_course_pct": 0.45,
+        "api_failure_pct": 0.25,
+        "amr_base_rate": 0.35,
+        "quality_tested_pct": 0.02,
+    },
+    "hospital_referral": {
+        "setting_probs": {"district_hospital": 0.35, "regional_hospital": 0.25,
+                          "tertiary_hospital": 0.20, "private_clinic": 0.20},
+        "antibiotic_probs": {"amoxicillin": 0.15, "ceftriaxone": 0.20, "ciprofloxacin": 0.15,
+                             "gentamicin": 0.10, "metronidazole": 0.10, "azithromycin": 0.10,
+                             "meropenem": 0.05, "vancomycin": 0.05, "cotrimoxazole": 0.10},
+        "sf_prevalence": 0.12,
+        "no_prescription_pct": 0.15,
+        "incomplete_course_pct": 0.20,
+        "api_failure_pct": 0.10,
+        "amr_base_rate": 0.40,
+        "quality_tested_pct": 0.08,
+    },
+    "cross_border_unregulated": {
+        "setting_probs": {"border_market": 0.35, "unlicensed_shop": 0.25,
+                          "mobile_vendor": 0.20, "online_seller": 0.20},
+        "antibiotic_probs": {"amoxicillin": 0.20, "cotrimoxazole": 0.20, "tetracycline": 0.15,
+                             "ciprofloxacin": 0.12, "ampicillin": 0.10, "chloramphenicol": 0.08,
+                             "doxycycline": 0.08, "erythromycin": 0.07},
+        "sf_prevalence": 0.42,
+        "no_prescription_pct": 0.85,
+        "incomplete_course_pct": 0.55,
+        "api_failure_pct": 0.35,
+        "amr_base_rate": 0.45,
+        "quality_tested_pct": 0.01,
+    },
+}
+
+SCENARIO_FILES = {
+    "community_otc_access": "antibiotic_community.csv",
+    "hospital_referral": "antibiotic_hospital.csv",
+    "cross_border_unregulated": "antibiotic_cross_border.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(25, 18), 0, 80))
+        sex = rng.choice(["male", "female"], p=[0.48, 0.52])
+
+        antibiotic = _choice(rng, params["antibiotic_probs"])
+        indication = rng.choice(["respiratory_infection", "UTI", "skin_wound",
+                                 "diarrhoea", "STI", "surgical_prophylaxis", "other"],
+                                p=[0.25, 0.15, 0.15, 0.15, 0.10, 0.10, 0.10])
+        no_prescription = int(rng.random() < params["no_prescription_pct"])
+        self_medication = int(no_prescription and rng.random() < 0.70)
+        incomplete_course = int(rng.random() < params["incomplete_course_pct"])
+        dose_subtherapeutic = int(rng.random() < 0.15)
+
+        manufacturer = rng.choice(["local_generic", "indian_generic", "chinese_generic",
+                                    "who_prequalified", "branded_originator", "unknown"],
+                                   p=[0.15, 0.30, 0.20, 0.15, 0.10, 0.10])
+
+        # Quality
+        is_sf = int(rng.random() < params["sf_prevalence"])
+        is_falsified = int(is_sf and rng.random() < 0.30)
+        is_substandard = int(is_sf and not is_falsified)
+
+        if is_falsified:
+            api_pct = float(np.clip(rng.normal(10, 15), 0, 40))
+        elif is_substandard:
+            api_pct = float(np.clip(rng.normal(60, 15), 20, 84))
+        else:
+            api_pct = float(np.clip(rng.normal(97, 4), 85, 115))
+
+        api_failure = int(api_pct < 85)
+        dissolution_failure = int(is_sf and rng.random() < 0.40)
+        contamination_detected = int(is_sf and rng.random() < 0.05)
+        wrong_ingredient = int(is_falsified and rng.random() < 0.15)
+
+        # AMR outcomes
+        sf_amr_mult = 1.8 if is_sf else 1.0
+        incomplete_mult = 1.5 if incomplete_course else 1.0
+        amr_detected = int(rng.random() < np.clip(
+            params["amr_base_rate"] * sf_amr_mult * incomplete_mult, 0, 0.80))
+        esbl_producer = int(amr_detected and rng.random() < 0.25)
+        mrsa = int(amr_detected and rng.random() < 0.10)
+        mdr = int(amr_detected and rng.random() < 0.15)
+
+        # Clinical outcomes
+        treatment_failure = int(rng.random() < np.clip(
+            0.10 * sf_amr_mult * incomplete_mult, 0, 0.40))
+        hospitalisation = int(treatment_failure and rng.random() < 0.20)
+        sepsis = int(hospitalisation and rng.random() < 0.15)
+        death = int(sepsis and rng.random() < 0.10)
+        adr = int(contamination_detected or (is_sf and rng.random() < 0.03))
+
+        # Surveillance
+        quality_tested = int(rng.random() < params["quality_tested_pct"])
+        culture_sensitivity_done = int(setting in ("district_hospital", "regional_hospital",
+                                                    "tertiary_hospital") and rng.random() < 0.15)
+        amr_reported = int(culture_sensitivity_done and amr_detected and rng.random() < 0.50)
+        antibiogram_available = int(culture_sensitivity_done and rng.random() < 0.30)
+
+        any_adverse = int(treatment_failure or adr or death)
+
+        record = {
+            "record_id": f"{name[:3].upper()}-{idx:05d}",
+            "scenario": name,
+            "year": year,
+            "setting": setting,
+            "age": age,
+            "sex": sex,
+            "antibiotic": antibiotic,
+            "indication": indication,
+            "no_prescription": no_prescription,
+            "self_medication": self_medication,
+            "incomplete_course": incomplete_course,
+            "manufacturer": manufacturer,
+            "is_substandard_falsified": is_sf,
+            "is_falsified": is_falsified,
+            "is_substandard": is_substandard,
+            "api_pct_label": round(api_pct, 1),
+            "api_failure": api_failure,
+            "dissolution_failure": dissolution_failure,
+            "wrong_ingredient": wrong_ingredient,
+            "amr_detected": amr_detected,
+            "esbl_producer": esbl_producer,
+            "mrsa": mrsa,
+            "mdr": mdr,
+            "treatment_failure": treatment_failure,
+            "hospitalisation": hospitalisation,
+            "sepsis": sepsis,
+            "death": death,
+            "adr": adr,
+            "quality_tested": quality_tested,
+            "culture_sensitivity_done": culture_sensitivity_done,
+            "amr_reported": amr_reported,
+            "any_adverse": any_adverse,
+        }
+        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 +1,90 @@
+"""Validate synthetic antibiotic quality & AMR acceleration dataset."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import matplotlib.pyplot as plt
+import pandas as pd
+
+SCENARIO_FILES = {
+    "community_otc_access": "antibiotic_community.csv",
+    "hospital_referral": "antibiotic_hospital.csv",
+    "cross_border_unregulated": "antibiotic_cross_border.csv",
+}
+
+COLORS = {"community_otc_access": "#e6550d", "hospital_referral": "#756bb1", "cross_border_unregulated": "#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["api_pct_label"], bins=50, alpha=0.5, color=COLORS[s], label=s, range=(0, 120))
+    axes[0].axvline(85, color="red", ls="--", lw=1, label="85% threshold")
+    axes[0].set_title("API Content (% of label)")
+    axes[0].legend(fontsize=6)
+
+    sf_cols = ["is_substandard_falsified", "is_falsified", "is_substandard"]
+    sf = df.groupby("scenario")[sf_cols].mean() * 100
+    sf.plot(kind="bar", ax=axes[1])
+    axes[1].set_title("SF Prevalence (%)")
+    axes[1].legend(fontsize=7)
+
+    abx = df.groupby(["scenario", "antibiotic"]).size().groupby(level=0).apply(lambda s: s / s.sum())
+    abx.unstack().plot(kind="bar", stacked=True, ax=axes[2])
+    axes[2].set_title("Antibiotic Distribution")
+    axes[2].legend(fontsize=4)
+
+    amr_cols = ["amr_detected", "esbl_producer", "mrsa", "mdr"]
+    amr = df.groupby("scenario")[amr_cols].mean() * 100
+    amr.plot(kind="bar", ax=axes[3])
+    axes[3].set_title("AMR Outcomes (%)")
+    axes[3].legend(fontsize=7)
+
+    use_cols = ["no_prescription", "self_medication", "incomplete_course"]
+    use = df.groupby("scenario")[use_cols].mean() * 100
+    use.plot(kind="bar", ax=axes[4])
+    axes[4].set_title("Antibiotic Use Patterns (%)")
+    axes[4].legend(fontsize=7)
+
+    out_cols = ["treatment_failure", "hospitalisation", "sepsis", "death"]
+    out = df.groupby("scenario")[out_cols].mean() * 100
+    out.plot(kind="bar", ax=axes[5])
+    axes[5].set_title("Clinical Outcomes (%)")
+    axes[5].legend(fontsize=7)
+
+    src = df.groupby(["scenario", "manufacturer"]).size().groupby(level=0).apply(lambda s: s / s.sum())
+    src.unstack().plot(kind="bar", stacked=True, ax=axes[6])
+    axes[6].set_title("Manufacturer Origin")
+    axes[6].legend(fontsize=5)
+
+    qa_cols = ["quality_tested", "culture_sensitivity_done", "amr_reported"]
+    qa = df.groupby("scenario")[qa_cols].mean() * 100
+    qa.plot(kind="bar", ax=axes[7])
+    axes[7].set_title("Surveillance & Testing (%)")
+    axes[7].legend(fontsize=7)
+
+    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()