Inspiration
We wanted our project to have real implications for Philadelphia, and Indego is Philly's official bike share program. The program has a published equity mandate to expand access in low-income communities. We wanted to build something that actually generated real, data-backed proposals that Indego and city planners could act on.
What it does
GoIndego identifies the highest-priority locations for new Indego bike share stations in Philadelphia. It visualizes ridership flows across the city, highlights stations that are over- or under-utilized, and overlays census and employment datasets to compute a custom gap score: a composite metric measuring how underserved a neighborhood is relative to its need. From there, Gemini AI automatically writes expansion proposals for each high-priority zone that are ready to submit to Indego.
How we built it
We pulled together our data sources from the census, LEHD, and Indego directly. We engineered a gap score metric that weights population density, income level, employment in the area, and existing station coverage. The frontend renders multiple interactive map layers showing ridership flows and scored zones. Gemini AI takes the zone-level data and generates expansion proposals.
Challenges we ran into
Merging datasets from different sources with different geographies was the hardest technical problem. Getting the gap score to feel meaningful required a lot of iteration on the weighting formula.
Accomplishments that we're proud of
The gap score is something we invented. We're proud that it surfaces neighborhoods that raw ridership data alone would miss entirely. We're also proud that the Gemini-generated proposals are arguments grounded in real data.
What we learned
We learned how to thoughtfully combine datasets with different methodologies and how to use AI to our advantage during our development. We also learned that framing a technical tool as a public accountability mechanism makes it far more compelling to a broader audience.
What's next for GoIndego
The gap score is designed to improve with more data, so adding more data about ridership would sharpen the recommendations significantly. We'd also like to build a public-facing version so Philadelphia residents can see the gaps in their own neighborhoods and add pressure on Indego to act. Longer term, the framework is replicable. Any city with open bike share data and an equity mandate could run the same analysis.
Log in or sign up for Devpost to join the conversation.