City-scale digital twin of 's-Hertogenbosch

City data lived in separate systems with different formats, refresh rates and keys. Answering one question about one street meant opening four tools and reconciling them by hand.

• ✦Event-driven ingestion with Kafka for live feeds and scheduled batch loads for registry data
• ✦Bronze → silver → gold lakehouse on Spark and Delta Lake, with schema enforcement and data quality checks
• ✦A geospatial city model that joins every source to a building ID
• ✦Low-latency serving APIs feeding a browser-based 3D viewer (CesiumJS)

Everything about a building on one screen, refreshed live. Shortlisted for the KIVI Award, and the platform the forecasting and AI work runs on.