App Logo
Decision Process
Using Google Stitch

About the Project

Inspiration

As students, we often book cabs late at night, after events, or for airport runs. We noticed we were frequently paying extra due to surge pricing or because our pickup spot wasn’t ideal. We wanted a tool that would make ride-hailing smarter: helping us decide whether to book now, wait a few minutes, or walk a block — without violating provider rules.

What We Built

We created Drift, a mobile companion app that:

Compares real-time fares from Uber, Lyft, and taxis.
Suggests smarter strategies:

Book Now (instant ride)
Wait & Save (predicted price drop if you wait a few minutes)
Walk Nearby (cheaper fares one block away)
- Lets you book directly through provider deep links.
- Tracks your ride history and shows total savings across trips.

How We Built It

Frontend: Designed with Google Stitch (Material 3 UI, scrollable cards, modern auth flow).
Backend / Data:
- Used mocked APIs and historical data to simulate surge pricing.
- Designed a lightweight rules engine to predict Wait & Save and Walk Nearby.
Integrations: Uber and Lyft deep links for booking.
Extras: A savings dashboard with total and per-ride insights.

What We Learned

How to integrate deep links with third-party ride apps.
Why user experience (UX) is as important as the backend logic — users want to see why they saved.
How to work with platform constraints and stay compliant with Uber/Lyft ToS.
Basics of modeling savings mathematically:

Inline example: The savings for a ride are defined as ( S = f_b - f_c ), where ( f_b ) is the baseline fare and ( f_c ) is the chosen fare.

Display example:

$$ S = f_b - f_c $$

Challenges

Data access: Live APIs are gated, so we simulated fare patterns using public datasets and heuristics.
Time: Balancing feature completeness with polished UI in a short hackathon sprint.
Consistency: Designing one flow that supports Book Now, Wait, and Walk without overwhelming the user.

What’s Next

Machine learning models to forecast short-term fare changes.
Event and weather data integrations for better predictions.
Push notifications: “Wait 5 minutes and save $4.”
Personalization: adapting recommendations to preferences (e.g., max wait time, max walking distance).