ncollo/Carbon-Trace-Kenya

AI-powered transport emission disclosure platform for Kenyan institutions. Automatically ingests fleet data, fuel records and travel receipts, applies GHG Protocol methodology with Kenya-specific calibrations, and generates NSE-compliant carbon reports in hours, giving EPRA a national emission baseline for evidence-based climate policy. # Carbon-Trace-Kenya AI-powered transport emission disclosure platform for Kenyan institutions. Features - Document ingestion (PDF, receipts, exports) - Cross-source reconciliation & anomaly detection - GHG Protocol calculation engine with Kenya adjustments - NSE-compliant XBRL + PDF disclosure generation - Privacy-preserving federated sector analytics (PySyft) Project structure (backend) Modules M1 | AI Document Intelligence & Data Ingestion - ingestion/ — document parsing & normalization - api/routers/upload.py — `POST /upload` M2 | Cross-Source Reconciliation & Anomaly Detection - reconciliation/ — anomaly pipelines and rule engine - api/routers/anomalies.py — `GET /anomalies` M3 | GHG Protocol Emission Calculation Engine (Core) - ghg_engine/ — emission factor registry & calculators - api/routers/calculate.py — `POST /calculate/{company_id}` M4 | NSE-Compliant XBRL Report & PDF Generation - reporting/ — narrative generator, XBRL and PDF builders - api/routers/reports.py — `GET /reports/{id}/pdf` · `GET /reports/{id}/xbrl` M5 | EPRA Sector Analytics & Federated Learning - federated/ — PySyft aggregator, consent and anonymiser - api/routers/epra/ — sector analytics endpoints Core shared infra - db/ — SQLAlchemy models & session pool - api/main.py — FastAPI app & router registration - config.py, dependencies.py, middleware.py — app wiring Recommended build order 1. `api/main.py` — FastAPI skeleton 2. `db/models.py` — define DB tables 3. `ghg_engine/emission_factor_registry.py` — pin factors 4. `ghg_engine/scope1_calculator.py` — core engine 5. `reconciliation/isolation_forest.py` — anomaly detection 6. `reporting/xbrl_builder.py` — XBRL output 7. `federated/fl_aggregator.py` — federated aggregation Contributors Timothy Gitau Muchiri — Team Lead & Backend AI Engineer EmitIQ Team — EPRA Hackathon 2026 Background processing & object storage - Configure S3 and Redis in `.env` or environment variables (`use_s3`, `s3_bucket`, `s3_access_key`, `s3_ …