Inspiration

Singapore has around 825,300 residents aged 65+ (21.4% of the population), and they're the most vulnerable to heat stress, but there is no real-time way to see who's actually at risk. We want to turn live heat and vulnerability data into risk levels governments can act on and residents can use to stay safe.

How we plan to build it

Ingest live NEA weather (temperature, humidity, WBGT), PM2.5, rainfall, and real eldercare/population/boundary data from data.gov.sg every 5 minutes via GitHub Actions, triggered by cron-job.org, loaded into Delta tables in Unity Catalog through Lakeflow Jobs.

We will layer on four models, which are Risk, Forecast, Site Suggestion, and Simulation, on top of the data, delivered through an AI/BI Dashboard for the live heat map and risk index, and a Databricks App for simulating new cooling-site placement.