There Are No Clean Maps

We talk about "putting data on a map" as if geography were stable. Crisp boundaries. Clear hierarchies. Agreed definitions. In practice, none of that exists.

Every map is a reconciliation. Between data sources collected at different times, administrative boundaries that have been redrawn, spatial units that refuse to nest cleanly into each other, and datasets whose geographic provenance is simply unknown.

This talk is about that invisible work. Drawing from real spatial analysis across village, pincode, hex and district-level data in India, Supriya walks through five situations where geographies break down: hierarchies that are conceptually nested but spatially inconsistent, parallel units like districts and pincodes that neither align nor exclude each other, granularity that has to be constructed from scratch, spatial logics that fundamentally conflict (raster grids versus hex systems), and survey data that arrives at the district level without clarity on which boundary definition was used.

For each case, the talk surfaces the analytical decisions involved, what gets assumed, what gets lost, and how much uncertainty gets quietly carried forward into the final visualisation.

The argument is not that clean maps are impossible. It is that spatial visualisation begins long before design, in decisions most practitioners never make visible. The question is not whether trade-offs exist. It is whether we acknowledge and communicate them.

If you work with geographic data at any level of granularity, this talk will change how you think about the map you are building.