I just stumbled across this. I’m the author of Geo::DNA. It’s so cool to see it in use like this!

One tip: what you say about Geo::DNA code prefixes being the same means that the locations are close is totally true, and as I mention on the site, it can sometimes be tricky when your datapoints cross boundaries of the bounding boxes – but you can easily get around that using the neighbours_within_radius () function to get all the codes within an area – any codes with those radius codes as their prefixes *are* location nearby.

There’s an interactive JS demo of this feature on http://www.geodna.org if you hadn’t already seen it.

Cheers and thanks for using it!

Kyle