Problem Statement

I rely on maps every day. Basic map apps are frustrating: they don’t understand natural language well, they don’t remind you of places based on your history, and they don’t auto-optimise your stops. You type “gym” or “grab coffee on the way” and get nothing useful. You already know you often stop at the same pharmacy on Thursday evenings — but your map never suggests it. I wanted navigation that actually listens. You say what you need in plain language; the app detects stops, finds them along your route, and when you upload your timeline, suggests places you usually visit on that trip.

Solution

Power Navigator is a route planner that helps you plan route based on natural language queries, it helps create an effortless experience by providing following features:

  • Detects stops in user's query.
  • Find locations based on stops along the route (i.e. source -> destination)
  • Discards irrelevant suggestion a normal map would suggest.
  • Recognize useful patterns in users timeline history.
  • Suggest locations from identified patterns by finding their relevance to current route as well as metrics like day and time of the day.

How this leverages Nova 2 Lite:

  • Detects stops in the user’s query — Nova turns the user’s message into a list of stop types (e.g. gym, pharmacy)

  • Find locations based on stops along the route — Nova supplies the stop types and extracts place names from turn-by-turn text so the app knows where along the route to run place search.

  • Recognize useful patterns in the user’s timeline — Nova reads the timeline (by day) and returns patterns (e.g. “On Thursdays you usually go from A to B via C”).

  • Suggest locations from patterns — Nova decides which patterns match the current route, day, and time of day, then extracts place names from those patterns and suggest them to user.

  • Decision-making — Nova handles the semantic decisions (what stops, what’s a duplicate, which patterns match, which places to pull from patterns)

Built With

Share this project:

Updates