Yelp API + Mapbox Isochrones for Autonomous Robot Deliveries
Built a Python tool that pulled all restaurants in a radius from the Yelp API, then used Mapbox isochrones to map which ones the REV-1 delivery robots could actually reach reliably given hills and speed. Surfaced 200+ new leads.
Context
Refraction AI operates autonomous delivery robots (the REV-1) for last-mile logistics. As a Product Associate on the go-to-market side, I needed to identify commercial partners near our operational routes — but the existing tooling couldn't answer the real question: "can the robot actually get there reliably?"
What I Built
A two-step Python pipeline:
- →Step 1 — Yelp API: pulled all restaurants in a radius around our office (and the 78704 zip code when we expanded), including contact info so the sales team could start dialing immediately.
- →Step 2 — Mapbox Isochrone API: instead of a naive distance circle, generated isochrones — travel-time contours that account for hills and terrain. Flat ground at 15 mph is fine for the REV-1. A steep hill isn't. The isochrone told us which restaurants were actually reachable at robot speed with reliable delivery time.
The output fed the sales team a ranked list of qualified prospects with contact info — and a visual map of our real operational footprint.
Why It Mattered
Before this: the sales team manually searched Google Maps, guessed at which places were reachable, and spent hours per lead. After: a filtered list of pre-qualified prospects with phone numbers, ready to call.
Results
- →Surfaced 200+ new potential clients in the first deployment
- →Replaced manual Google Maps research with API-driven prospecting
- →Improved customer retention rates by 25% through product enhancements based on user feedback