Using Hexagons to Address Risk

Addresses are the way people think about place - from ordering pizza to getting an insurance quote - it all starts with an address. But, address data is big. There are 31 million addresses in the UK, and 160 million in the United States. At Addresscloud we need to be able to lookup an address and place it on a map in real-time. Complicating the problem is the hundreds of changes published every day, as addresses are created, moved and deleted from national databases. Maintaining the relationship between an address and its surroundings is a "big" spatial data challenge, and one we spend a lot of time trying to optimise.

We've recently started using the H3 geospatial index to map relationships between addressess and geographic data. Using a consistent heirarchical index is helping reduce the problem space, makes our queries faster, and allows spatial aggregation of information on-the-fly. However, rather than write an extended blog post on the topic I'm happy to be able to include my presentation from the recent 2020 Spatial Data Science Conference!

It was a real privalge to be invited to speak at SDSC20 and be part of such an engaging line-up of speakers. My thanks to the team at Carto who did a great job of running the event. You can watch all the presentation videos online at the conference website.