Add dataset files

README.md

@@ -0,0 +1,77 @@
+---
+license: cc-by-4.0
+task_categories:
+  - tabular-classification
+  - tabular-regression
+language:
+  - en
+tags:
+  - healthcare
+  - medicine-quality
+  - supply-chain
+  - track-and-trace
+  - serialization
+  - diversion
+  - GS1
+  - pharmaceutical
+  - sub-saharan-africa
+  - lmic
+pretty_name: "Supply Chain Integrity & Track-and-Trace (Serialization, Diversion, SF Entry)"
+size_categories:
+  - 10K<n<100K
+configs:
+  - config_name: serialized_track_trace
+    data_files: data/sct_serialized_track_trace.csv
+  - config_name: partial_visibility
+    data_files: data/sct_partial_visibility.csv
+    default: true
+  - config_name: opaque_uncontrolled
+    data_files: data/sct_opaque_uncontrolled.csv
+---
+
+# Supply Chain Integrity & Track-and-Trace Dataset
+
+## Abstract
+
+**30,000 simulated medicine shipment observations** (10,000 per scenario) across three supply chain visibility levels. Variables include serialization, authentication, temperature monitoring, GPS tracking, intermediary count, diversion, tampering, SF product entry, and recall. Three scenarios: serialized (90% serialized, 3% SF), partial visibility (16% serialized, 23% SF), opaque (1% serialized, 69% SF).
+
+**This dataset is entirely simulated. It must not be used for supply chain or regulatory decisions.**
+
+## Validation
+
+<p align="center">
+  <img src="validation_report.png" alt="Validation Report" width="100%">
+</p>
+
+## Usage
+
+```python
+from datasets import load_dataset
+dataset = load_dataset("electricsheepafrica/supply-chain-track-trace", "partial_visibility")
+df = dataset["train"].to_pandas()
+print(df.groupby('serialized')['sf_product_detected'].mean())
+```
+
+## References
+
+1. WHO (2024). Complex supply chains and SF medicines.
+2. GS1 Healthcare. Pharmaceutical serialization standards.
+3. AU/AMRH. Track-and-trace pilots in Africa.
+4. EU Falsified Medicines Directive.
+5. USAID GHSC-PSM. Supply chain visibility tools.
+
+## Citation
+
+```bibtex
+@dataset{esa_sct_2025,
+  title   = {Supply Chain Integrity and Track-and-Trace Dataset},
+  author  = {{Electric Sheep Africa}},
+  year    = {2025},
+  publisher = {Hugging Face},
+  url     = {https://huggingface.co/datasets/electricsheepafrica/supply-chain-track-trace}
+}
+```
+
+## License
+
+[CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)

generate_dataset.py

@@ -0,0 +1,195 @@
+#!/usr/bin/env python3
+"""
+Literature-Informed Supply Chain Integrity & Track-and-Trace Dataset
+=====================================================================
+
+Each record = ONE medicine shipment/consignment assessed for integrity.
+
+Sources (v2.0):
+[1] WHO (2024). Complex supply chains with multiple intermediaries
+    increase risk of product tampering and substitution.
+[2] GS1 Healthcare. Global standards for pharmaceutical serialization
+    and track-and-trace. GTIN, SSCC, 2D barcodes.
+[3] African Union/AMRH. Track-and-trace pilots in East Africa,
+    Nigeria, Ghana. Mobile verification systems (mPedigree, Sproxil).
+[4] EU Falsified Medicines Directive. End-to-end serialization model.
+    Only ~5 African countries have mandatory serialization.
+[5] USAID GHSC-PSM. Supply chain visibility tools for donor-funded
+    medicines. Temperature monitoring, GPS tracking.
+"""
+
+import numpy as np
+import pandas as pd
+import argparse
+import os
+
+SUPPLY_CHAIN_STAGES = [
+    'manufacturer', 'primary_wholesaler', 'regional_distributor',
+    'clearing_agent_port', 'central_medical_store', 'regional_store',
+    'district_store', 'health_facility', 'retail_pharmacy',
+]
+
+PRODUCT_TYPES = [
+    ('essential_medicine', 'tablet', 0.35),
+    ('antimalarial_ACT', 'tablet', 0.15),
+    ('antibiotic', 'capsule', 0.12),
+    ('ARV', 'tablet', 0.08),
+    ('vaccine', 'vial', 0.08),
+    ('injectable', 'ampoule', 0.07),
+    ('oral_liquid', 'bottle', 0.05),
+    ('controlled_substance', 'tablet', 0.03),
+    ('medical_device', 'unit', 0.04),
+    ('biological', 'vial', 0.03),
+]
+
+SCENARIOS = {
+    'serialized_track_trace': {
+        'system': 'full_serialization',
+        'serialization_coverage': 0.90,
+        'authentication_rate': 0.85,
+        'diversion_rate': 0.02,
+        'tampering_detected': 0.01,
+        'temperature_monitoring': 0.80,
+        'GPS_tracking': 0.70,
+        'sf_entry_rate': 0.03,
+        'transit_days_mean': 15,
+        'intermediary_count_mean': 3,
+    },
+    'partial_visibility': {
+        'system': 'partial_visibility',
+        'serialization_coverage': 0.15,
+        'authentication_rate': 0.20,
+        'diversion_rate': 0.08,
+        'tampering_detected': 0.04,
+        'temperature_monitoring': 0.25,
+        'GPS_tracking': 0.15,
+        'sf_entry_rate': 0.12,
+        'transit_days_mean': 30,
+        'intermediary_count_mean': 5,
+    },
+    'opaque_uncontrolled': {
+        'system': 'opaque',
+        'serialization_coverage': 0.01,
+        'authentication_rate': 0.02,
+        'diversion_rate': 0.20,
+        'tampering_detected': 0.10,
+        'temperature_monitoring': 0.03,
+        'GPS_tracking': 0.02,
+        'sf_entry_rate': 0.30,
+        'transit_days_mean': 60,
+        'intermediary_count_mean': 8,
+    },
+}
+
+
+def generate_dataset(n=10000, seed=42, scenario='partial_visibility'):
+    rng = np.random.default_rng(seed)
+    sc = SCENARIOS[scenario]
+    records = []
+    n_prod = len(PRODUCT_TYPES)
+
+    for idx in range(n):
+        rec = {'id': idx + 1}
+        rec['system'] = sc['system']
+        rec['shipment_id'] = f"SCT_{rng.integers(1, 5000):05d}"
+
+        prod = PRODUCT_TYPES[rng.choice(n_prod, p=[p[2] for p in PRODUCT_TYPES])]
+        rec['product_type'] = prod[0]
+        rec['dosage_form'] = prod[1]
+
+        rec['origin_country'] = rng.choice(
+            ['India', 'China', 'Europe', 'local_SSA', 'USA', 'South_Africa'],
+            p=[0.40, 0.18, 0.10, 0.12, 0.05, 0.15])
+        rec['destination_country'] = rng.choice(
+            ['Nigeria', 'Kenya', 'Tanzania', 'Ghana', 'Ethiopia',
+             'DRC', 'Uganda', 'Cameroon', 'Senegal', 'Mozambique'],
+            p=[0.15, 0.12, 0.10, 0.10, 0.10, 0.08, 0.08, 0.08, 0.10, 0.09])
+
+        rec['intermediary_count'] = max(1, int(rng.poisson(sc['intermediary_count_mean'])))
+        rec['transit_days'] = max(1, int(rng.exponential(sc['transit_days_mean'] * 0.6)))
+        rec['storage_handoffs'] = max(1, min(rec['intermediary_count'], int(rng.poisson(3))))
+
+        rec['serialized'] = 1 if rng.random() < sc['serialization_coverage'] else 0
+        rec['barcode_2D'] = 1 if rec['serialized'] and rng.random() < 0.85 else 0
+        rec['GTIN_assigned'] = 1 if rec['serialized'] else (1 if rng.random() < 0.10 else 0)
+        rec['authentication_scanned'] = 1 if rng.random() < sc['authentication_rate'] else 0
+        rec['mobile_verification_used'] = 1 if rng.random() < (
+            sc['authentication_rate'] * 0.5) else 0
+
+        rec['temperature_monitored'] = 1 if rng.random() < sc['temperature_monitoring'] else 0
+        rec['temperature_excursion'] = 0
+        if not rec['temperature_monitored'] or rng.random() < (0.05 if scenario == 'serialized_track_trace' else (0.15 if scenario == 'partial_visibility' else 0.40)):
+            rec['temperature_excursion'] = 1 if rng.random() < 0.30 else 0
+        rec['GPS_tracked'] = 1 if rng.random() < sc['GPS_tracking'] else 0
+
+        rec['customs_cleared'] = 1 if rng.random() < (0.98 if scenario == 'serialized_track_trace' else (0.85 if scenario == 'partial_visibility' else 0.60)) else 0
+        rec['import_permit_valid'] = 1 if rng.random() < (0.95 if scenario == 'serialized_track_trace' else (0.70 if scenario == 'partial_visibility' else 0.30)) else 0
+        rec['WHO_prequalified_product'] = 1 if rng.random() < (0.55 if scenario == 'serialized_track_trace' else (0.25 if scenario == 'partial_visibility' else 0.05)) else 0
+
+        # Integrity issues
+        rec['diversion_detected'] = 1 if rng.random() < sc['diversion_rate'] else 0
+        rec['tampering_evidence'] = 1 if rng.random() < sc['tampering_detected'] else 0
+        rec['packaging_breach'] = 1 if rng.random() < (sc['tampering_detected'] * 1.5) else 0
+
+        # SF entry into supply chain
+        base_sf = sc['sf_entry_rate']
+        if not rec['serialized']:
+            base_sf *= 1.5
+        if rec['intermediary_count'] > 5:
+            base_sf *= 1.3
+        if not rec['customs_cleared']:
+            base_sf *= 2.0
+        if rec['diversion_detected']:
+            base_sf *= 2.0
+        base_sf = np.clip(base_sf, 0.005, 0.80)
+
+        rec['sf_product_detected'] = 1 if rng.random() < base_sf else 0
+        rec['sf_entry_point'] = 'none'
+        if rec['sf_product_detected']:
+            rec['sf_entry_point'] = rng.choice(
+                ['manufacturer_level', 'wholesaler', 'port_customs',
+                 'regional_distribution', 'last_mile', 'retail'],
+                p=[0.15, 0.20, 0.15, 0.20, 0.15, 0.15])
+
+        rec['recall_initiated'] = 0
+        if rec['sf_product_detected'] and rng.random() < (
+            0.60 if scenario == 'serialized_track_trace' else
+            (0.15 if scenario == 'partial_visibility' else 0.02)):
+            rec['recall_initiated'] = 1
+
+        rec['batch_quantity'] = max(100, int(rng.lognormal(8, 1)))
+        rec['value_usd'] = round(rec['batch_quantity'] * max(0.01, rng.lognormal(
+            np.log(0.50), 0.8)), 2)
+
+        rec['year'] = rng.choice([2020, 2021, 2022, 2023, 2024],
+                                 p=[0.10, 0.15, 0.20, 0.25, 0.30])
+
+        records.append(rec)
+
+    df = pd.DataFrame(records)
+    print(f"\n{'='*65}")
+    print(f"Supply Chain — {scenario} (n={n}, seed={seed})")
+    print(f"{'='*65}")
+    print(f"  System: {sc['system']}")
+    print(f"  Serialized: {df['serialized'].mean()*100:.1f}%")
+    print(f"  SF detected: {df['sf_product_detected'].mean()*100:.1f}%")
+    print(f"  Diversion: {df['diversion_detected'].mean()*100:.1f}%")
+    print(f"  Avg intermediaries: {df['intermediary_count'].mean():.1f}")
+    return df
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--all-scenarios', action='store_true')
+    parser.add_argument('--n', type=int, default=10000)
+    parser.add_argument('--seed', type=int, default=42)
+    args = parser.parse_args()
+    os.makedirs('data', exist_ok=True)
+    if args.all_scenarios:
+        for sc in SCENARIOS:
+            df = generate_dataset(n=args.n, seed=args.seed, scenario=sc)
+            df.to_csv(os.path.join('data', f'sct_{sc}.csv'), index=False)
+            print(f"  -> Saved\n")
+    else:
+        df = generate_dataset(n=args.n, seed=args.seed)
+        df.to_csv(os.path.join('data', 'sct_partial_visibility.csv'), index=False)

requirements.txt

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

validate_dataset.py

@@ -0,0 +1,106 @@
+#!/usr/bin/env python3
+"""Validation & Diagnostic Visualization for Supply Chain Integrity & Track-and-Trace Dataset."""
+
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import os
+
+SCENARIOS = ['serialized_track_trace', 'partial_visibility', 'opaque_uncontrolled']
+
+
+def load_scenarios(data_dir='data'):
+    dfs = {}
+    for sc in SCENARIOS:
+        path = os.path.join(data_dir, f'sct_{sc}.csv')
+        if os.path.exists(path):
+            dfs[sc] = pd.read_csv(path)
+    return dfs
+
+
+def make_report(dfs, output='validation_report.png'):
+    fig, axes = plt.subplots(4, 2, figsize=(16, 24))
+    fig.suptitle(
+        'Supply Chain Integrity & Track-and-Trace — Validation Report\n'
+        '(Serialized → Partial Visibility → Opaque)',
+        fontsize=15, fontweight='bold', y=0.99)
+    colors = ['#2ecc71', '#f39c12', '#e74c3c']
+    x = np.arange(len(SCENARIOS))
+    labels = ['Serialized', 'Partial', 'Opaque']
+
+    ax = axes[0, 0]
+    sf = [dfs[sc]['sf_product_detected'].mean()*100 for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, sf, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(sf):
+        ax.text(i, v+1, f'{v:.0f}%', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('SF Rate (%)'); ax.set_title('SF Product Detection Rate')
+
+    ax = axes[0, 1]
+    ser = [dfs[sc]['serialized'].mean()*100 for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, ser, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(ser):
+        ax.text(i, v+1, f'{v:.0f}%', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('Rate (%)'); ax.set_title('Serialization Coverage')
+
+    ax = axes[1, 0]
+    div = [dfs[sc]['diversion_detected'].mean()*100 for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, div, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(div):
+        ax.text(i, v+0.3, f'{v:.1f}%', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('Rate (%)'); ax.set_title('Diversion Detected')
+
+    ax = axes[1, 1]
+    interm = [dfs[sc]['intermediary_count'].mean() for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, interm, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(interm):
+        ax.text(i, v+0.1, f'{v:.1f}', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('Count'); ax.set_title('Average Intermediaries')
+
+    ax = axes[2, 0]
+    df = dfs.get('partial_visibility', list(dfs.values())[1])
+    ep = df[df['sf_product_detected']==1]['sf_entry_point'].value_counts()
+    if len(ep) > 0:
+        ax.barh(range(len(ep)), ep.values, color='#e74c3c', alpha=0.7)
+        ax.set_yticks(range(len(ep)))
+        ax.set_yticklabels([s.replace('_', ' ').title() for s in ep.index], fontsize=7)
+        ax.set_xlabel('Count')
+    ax.set_title('SF Entry Points (Partial Visibility)')
+
+    ax = axes[2, 1]
+    temp = [dfs[sc]['temperature_monitored'].mean()*100 for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, temp, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(temp):
+        ax.text(i, v+1, f'{v:.0f}%', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('Rate (%)'); ax.set_title('Temperature Monitoring')
+
+    ax = axes[3, 0]
+    transit = [dfs[sc]['transit_days'].median() for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, transit, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(transit):
+        ax.text(i, v+0.5, f'{v:.0f}d', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('Days'); ax.set_title('Median Transit Days')
+
+    ax = axes[3, 1]
+    recall = [dfs[sc]['recall_initiated'].mean()*100 for sc in SCENARIOS if sc in dfs]
+    ax.bar(x, recall, color=colors, alpha=0.8)
+    ax.set_xticks(x); ax.set_xticklabels(labels, fontsize=9)
+    for i, v in enumerate(recall):
+        ax.text(i, v+0.2, f'{v:.1f}%', ha='center', fontsize=10, fontweight='bold')
+    ax.set_ylabel('Rate (%)'); ax.set_title('Recall Initiated')
+
+    plt.tight_layout(rect=[0, 0, 1, 0.97])
+    plt.savefig(output, dpi=150, bbox_inches='tight')
+    print(f'Saved validation report to {output}')
+    plt.close()
+
+
+if __name__ == '__main__':
+    dfs = load_scenarios()
+    if dfs:
+        make_report(dfs)