Map Search Revamp (2023)

My drive for detailed user experiences and amazing interface details led me to this project, where we taken a considerable laggy and broken legacy map search experience and totally revamped it.

The problem

As a proptech, search by map is a core experience for the user who is looking for a new place to buy or to rent. The existing map codebase, besides being increasingly unmaintanable and hard to read, produced a laggy and buggy experience for the user, mainly affecting those with low end Android devices.

The solution

I lead a series of initiatives:

1. Reconstructed all the logic of the map search by migrating from MVVM to Flutter Bloc + Clean Architecture Pattern, improving significantly the performance and fixing all major bugs in the process.
2. Created a new look and feel for the pins that represent the prices of the houses on the map. Needed to create this renderization on Native Side (a fork of Google Maps Flutter + custom Swift and Kotlin code) because of performance constraints.
3. Rethink the search fetch logic, talking to Product and Design team to find a balance on when to refetch houses when map moved, how much is the appropriate density based on screen size and number of houses found. Implemented this both on client side and made the necessary adjustments on the backend (Python + Elastic Search with Geospatial logic)
4. Decision and separation of initiatives via series of feature toggles (using Firebase Remote Config) to make experiments and phased rollouts (around 10 experiments were launched in a timeframe of 6 months).

With all experiments compound, we reached a total of 15% in increase of conversion in offers from tenants and buyers who look for houses.