Automating aGIS pipeline
Ramped onto an unfamiliar geospatial codebase in weeks and took the manual steps out of map publishing in Bhugyan GIS.
Publishing a map meant a person doing the same sequence by hand every time: ingest the raster layers, publish the WMS, seed it, and unpublish when it changed. Slow, and easy to get wrong in a way you only notice once the map renders.
A two-month internship on a domain — geospatial systems — with its own vocabulary and its own failure modes. Getting anything shipped meant learning the codebase and the subject matter at the same time, through close and frequent coordination with a mentor and subject-matter experts.
What I built
- Step 01
- Step 02
- Step 03
- 01
Automated raster ingestion
Raster layer ingestion became automatic rather than hand-run, which is what brought end-to-end map render time down.
- 02
CLI and Docker pipelines
WMS publishing, seeding and unpublishing were wrapped into pipelines, so the sequence runs the same way every time instead of depending on who is at the keyboard.
- 03
Extending the automation framework
Added coordinate conversions, GIS annotations, and graph-based algorithms for shortest path and visibility analysis.