What PathPay Delivers
Recommendation Engine
- CSV ingestion & cleaning – Automatically merges multiple datasets and
filters to Seoul’s Gwangjin/Songpa stations.
- Kakao Local geocoding – Resolves every station address into a precise
latitude/longitude.
- Smart candidate selection – Keeps stations within ±1 km of the route to
minimize unnecessary API calls.
- Total cost calculator – Adds detour fuel and time value to the raw station
price to compute an “effective price per liter.”
Driver Experience
- Top 3 summary – Console report highlights name, price, detour impact, and
savings.
- Kakao Map deeplink – One-click link launches the recommended station in
Kakao Map.
- Folium visualization – Optional HTML map plots origin, destination, and
the leading stations.
Operations
- Configurable knobs – Adjust detour buffer, number of candidates, and time
value to match fleet policies.
- Extensible APIs – Core logic is packaged in modules that can be reused
from web services, schedulers, or notebooks.
- Documentation-first – Full setup, architecture, and API docs live on
ReadTheDocs.