The problem
Reviewing performance scores meant working through large amounts of data by hand. A dashboard had to make that data easy to filter and compare without slowing down as the data grew.
The approach
- Server state with TanStack Query: caching and background refresh keep data current without blocking the screen or refetching on every interaction.
- Large-table virtualisation: only the rows on screen are rendered, so tables stay smooth regardless of size.
- Reusable chart and table building blocks give every report the same look and behaviour.
- Environment-specific API configuration for Dev, UAT and Production.
- API (per environment)
- TanStack Query cache + background refresh
- Filters
- Virtualised tables and charts
Simplified data flow.
Results
- 80%less manual review time
- Responsiveunder heavy datasets
- Consistentreporting UX from shared primitives
This was client work under NDA. Diagrams are simplified redraws, and no client data or screenshots are shown.