GIS Map Platform & Performance Optimization (Locable)
Project Overview & Objective
- GIS platform developed as a service provider for project teams to consume
- Custom map layers with Mapbox-based style customization
- No-code experience: users customize widgets (keystat, filter, map controller, search location, presentation, report) to display on the map
- Publish map as URL for iframe embedding or install via npm component
- Role-based display control per user key
- Reduce point-loading latency to meet production deploy criteria
- Enable fast and stable map loading for users
Role & Responsibilities
- Investigated slow point loading with the team
- Designed and executed k6 load testing to measure baseline and validate results
- Re-designed database diagram and pre-processed points into .geojson, vector tiles by required type
- Implemented Redis caching for repeated user data to optimize response time
Tech Stack & Tools
MapboxGeoJSONk6RedisPostgreSQL
Key Challenges & Solutions
Challenges
- 8,000ms response time failed production deploy criteria
- Query from database plus post-processing — system slowed significantly with large data
- Source table was not normalized (synced from central without processing)
Solutions
- Re-designed database diagram and pre-processed points into .geojson stored by type
- Redis caching for repeated user data fetches
- Result: response time reduced to 1,200ms (85% reduction)