^ A demo of how to use a binary index I created with the following notebook:
This notebook takes a list of all the cities in the world with population > 1000, and then takes 65,536 samples distributed by population density. The samples are then packed into a tree using Uber’s H3 geo index, then packed and compressed into a binary index (around 252 KB) using CBOR and pako (published onto IPFS). With the index, you can pick any number from 0 to 65,535 and get a lat/long pair back corresponding to a populated location in the world weighted by population density. This might be useful for visualizations where you don’t know exactly where the data is coming from, but you’d like to assign a geo-location to it.