Logo Lanfrica

declerke/Kenya-Tenders-Pipeline

Domain:

natural language processingsocioeconomic

Record type:

software
Creator:
dec
Host:
Kenya procurement intelligence # πŸ›οΈ Kenya Tenders Pipeline: Government Procurement Intelligence **Kenya Tenders Pipeline** is a production-grade procurement intelligence system that scrapes live government and NGO tender listings from TendersKenya.co.ke, classifies every tender into one of seven industry sectors using spaCy NLP keyword matching, exposes the enriched dataset across a 4-page Streamlit dashboard with real-time urgency tracking, and orchestrates the full extract β†’ enrich β†’ transform β†’ test cycle as a daily Apache Airflow 3.0 DAG β€” giving any analyst, journalist, or supplier instant visibility into what Kenya's public procurement market is buying right now. | Metric | Value | |--------|-------| | Tender records | 315 (live scrape β€” 2025–2026 dates) | | Open tenders | 29 (genuine future deadlines) | | NLP-enriched rows | 280 / 315 across 7 sectors | | Airflow tasks | 5/5 SUCCESS (ingest β†’ NLP β†’ dbt run β†’ dbt test β†’ summary) | | dbt models | 5 (1 staging Β· 1 fact Β· 3 marts) | | dbt tests | 21/21 PASS | | Pytest tests | 48/48 PASS | | Dashboard pages | 4 | | Cost to run | $0 β€” open data + local stack only | --- ## 🎯 Project Goal Kenya's public procurement market processes billions of shillings in government and NGO contracts every year. Suppliers, civil society organisations, and journalists need to track what is being bought, by whom, and how urgently β€” but procurement data is scattered across dozens of entity-specific portals and aggregator sites, with no unified machine-readable feed. The official government portals (tenders.go.ke, ppra.go.ke) carry expired SSL certificates; HDX carries no relevant Kenya procurement datasets; OpenAFRICA's only Kenya tenders dataset is a static 2018–2019 CSV with no price data and no update schedule. Kenya Tenders Pipeline solves this by treating TendersKenya.co.ke as the data source β€” a live aggregator that consolidates government, county, NGO, and parastatal procurement notices into a single paginated listing with real 2025–2026 closing dates. …