I am tempted to import this data into my system and build a pivot table of building type (PRIM_OCC) by state.
I could then graph the data (pie chart, bar graph, etc) to show how the building type distribution (e.g. residential ratio per hospital) varies between the states.
An application that simulates shadows at a particular location. For real estate purposes, seeing what kind of sunshine you would get in your backyard when you purchase that house with a backyard in the north.
The geometry is just the building footprint. You could probably do some sort of estimate as to how likely the yard was to have sun, but it wouldn't enable doing anything detailed.
CoreLogic was doing something like this, but line of sight from a window and then be able to possibly identify the scene / view that can be had from that window.
I could then graph the data (pie chart, bar graph, etc) to show how the building type distribution (e.g. residential ratio per hospital) varies between the states.
https://m.youtube.com/watch?v=2ScBd-71OLQ
You can look at it in a slippy map: https://gis-fema.hub.arcgis.com/pages/usa-structures (a couple clicks required from there).
In my area it doesn't particularly identify garages well, so you probably can't spend that much time on bike sheds.
Some may need an assist: https://en.wiktionary.org/wiki/bikeshedding
I love to see a wry comment being taken seriously too and this one did not disappoint.