Large geospatial datasets are difficult to process and serve quickly. I built a fault-tolerant, asynchronous service that converts massive raster data into Cloud Optimized GeoTIFFs (COGs). I then designed a three-tier cache across raster, virtual-raster, and tile data to serve COGs in a triangular tiling format. The approach reduced web-rendering latency and memory use. The project taught me to optimize the full path from processing to delivery, not just the conversion step.
From large raster datasets to responsive maps
on 2026-11-01
From large raster datasets to responsive maps
