Clojurians Log v2
clojuredesign-podcast 2026-10-15
How do you keep outside data complexity from leaking into your application? How do you integrate changes in that data over time? In our latest episode, we take a look at pure data models that help us reason coherently about the outside world. https://clojuredesign.club/episode/120-data-vision/
Geojson for geographic boundries with government Covid data reports. Each local area for a covid report had different geographical boundaries than the boundaries defined by Geojson. Many areas in the scope of a report were either partial or crossed over geographical boundaries, so were challenging to align to create a visual map.
Oh wow, sounds pretty hairy. How were you able to align the boundaries? Did they remain the same across iterations?
I did a rough approximation when visualising the mismatched data sets (adding caveats). I started to create a geojson definition that matched the government data. Half way through that process the government changed boundaries again 😳 Eventually I created a simplified visual display (4 country boundaries, rather than ~100 counties - the UK has anywhere from 92 to 155 counties depending on what is considered a county 😂) I also enhanced how the numeric data was presented, so it was easier to consume.
ERP systems can spit out pretty horrible stuff IMO.
Hah, they changed the boundaries? Programming is just one assumption breaking after another.