last year i did a quick and dirty MVP for a friend who wanted to pitch his idea called locn to investors which was aiming to solve the last mile address issues. for this MVP i went into the rabbit hole of h3.

we decided to base the mvp on h3 for two reasons

1. it was a hierarchical spatial index system
2. it uses hexagons which comes handy in doing GIS analysis(equidistant from all adjacent shapes)

so when we started building the GIS center at meghalaya, we wanted to check if h3 can be really used for better analysis. so we used h3 based tiling of the state of meghalaya and used it to do anlaysis of school density and which locations need more attention. this gave a better visual representation of reality on the ground. we published a pre-computed layer in the dashboard which served as an entry point for the planning department and education department to run more realtime queries. we are planning to create more such precomputed layers for each of the classes that we are publishing and validating if it helps the departments do better analysis.

have you used h3 in any of your geo spatial analysis requirements? what are you thoughts? would you prefer h3 or do you prefer other spatial index?

i have added a link to the video, demo and other relevant links in the first comment.

youtube link - https://www.youtube.com/watch?v=E_M_kMmVhOw dashboard - https://meghalayacmdashboard.in/gisportal/apps/mapviewer/index.html?webmap=b4d717038bb5475fabe30d48fd1bd0ee plus codes - https://blog.google/products/maps/google-maps-101-giving-everyone-everywhere-an-address/ uber h3 - https://www.uber.com/en-GB/blog/h3/