From a spatial question to a working tool.

I designed and built the workspace and its supporting PostGIS and dbt data platform. The development version connects reviewed CV import, role search, saved applications, tailored CVs and cover letters, document exports and learning priorities. I designed the interface to distinguish a candidate’s stated experience from job requirements and missing information, then checked the connected journey using sample candidates.

Scope and limitsThe illustrated workspace is in development and uses sample data. Local journey checks do not establish production availability, complete job coverage or hiring outcomes. Matching supports shortlisting, not a calibrated probability of being hired.

I built the application with MapLibre, deck.gl, DuckDB-Wasm and CARTO. A versioned London Overture extract supplies GeoParquet for browser analysis and PMTiles for display. I used Valhalla walking polygons to define the analysis area and checked the local browser results against DuckDB. Other locations and sections use a separate CARTO provider.

Scope and limitsThe curated browser-data path covers London amenities. Other locations and sections depend on CARTO. Public deployment, production performance and wider coverage remain to be verified.

I built the data pipeline, mapping interface and browser search. Geographic coordinates determine position; language-model embeddings supply colour and similarity. Precomputed vectors, Arrow metadata and a web worker support search across the development subset. I also developed a separate study of geographic coverage across selected language editions.

Scope and limitsThe served map is an English development subset, and the language study samples selected editions rather than all of Wikipedia. Semantic similarity is a model-derived association, not proof of a factual connection between places.