ShelterGoWhere

Inspiration

Singapore is getting hotter, and elderly residents are among the most vulnerable to heat stress. Many still walk to the market, clinic or bus stop every day, yet heat is usually measured region by region, not route by route. We want to help planners answer: where should the next shelter or covered linkway go to protect the most elderly walkers?

What it does

Idea stage. This is what we plan to build.

ShelterGoWhere will:

• Map shade along footpaths in [planning area] across the day, using building footprints, building heights and sun position.
• Simulate elderly walking trips from homes to hawker centres, clinics and bus stops, weighted by where residents aged 65+ live.
• Test fixes before building them: a planner adds a covered linkway or trees on the map and sees how much heat exposure it removes.

How we built it

Planned approach, to be built on Databricks Free Edition:

• Load Singapore open data (building, population, weather and amenity datasets) into governed Delta tables.
• Compute shade and trip exposure in PySpark notebooks, saving each scenario as its own table.
• Present results in a simple map app, with Genie for plain-English questions.

Challenges we ran into

We haven't started building yet. The main challenges we expect are estimating shade from building data and the lack of real pedestrian trip data, which is why we simulate trips and will say so clearly.

Accomplishments that we're proud of

So far: choosing a focused problem, identifying the open datasets we need, and designing an end-to-end pipeline from raw data to a tool a planner can use.

What we learned

We're learning Databricks through the challenge's training session and will update this section during the build.

What's next for ShelterGoWhere

Build a working prototype for one planning area, then compare our shade estimates with real temperature readings.

Built With

• databricks
• nothing