French public data is plentiful and reliable, but spread across different organisations, formats and specialist terminology. EcoKwa provides a simpler entry point for finding an indicator, understanding what it represents and following its evolution.
What I built
Search and navigation across 16 indicators organised into 6 topics.
Pages combining the latest value, history, chart, table and methodology.
Comparison of compatible series over a shared period.
Automated imports from Insee, DREES and the French Pensions Advisory Council.
My role
I designed and developed EcoKwa independently, from data modelling and imports through to the front end, tests, deployment and operation of the application.
Technical decisions
Three challenges at the heart of the project
01
Importing heterogeneous sources
Data arrives through paginated JSON APIs and Excel files. Dedicated importers validate responses, normalise periods and store observations without creating duplicates when a job is replayed.
02
Making updates dependable
Imports run in the background with request limits, locks against concurrent runs and progressive retries for temporary failures. Every run keeps a history that can be inspected.
03
Preserving the meaning of the data
A missing value is never replaced with zero. Provisional, revised and partial statuses remain visible, and every chart has an accessible tabular alternative.
An indicator page brings together its latest value, evolution, origin and official source.Compatible series are aligned over a shared period without artificially filling missing observations.
Technical stack
A modern full-stack foundation serving the product
