Has anybody developed a spatial clustering method for identifying contiguous census block group clusters that minimize visits across cluster boundaries?

!channel has anybody developed a spatial clustering method for identifying contiguous census block group clusters that minimize visits across cluster boundaries?


This topic was automatically generated from Slack. You can find the original thread here.

I don’t know about clustering over borders based on visit patterns but simple spatial clustering is available in the fantastic package fixest in R as a post-estimation adjustment. I suppose you could also define a new variable with the conditions you have in mind (contiguous and traffic) and cluster on that explicitly.

Hi Derek! I think something like this could also be done using the skater function in R (described in this paper)

I do notwhat clusters you need, but you might take a look at our project here: US Mobility Maps — EndCoronavirus.org

Thanks all. Yeah I’m considering some of the methods you’ve shared. I think my interest is a very fundamental one – any of us who have considered the problem of spatial autocorrelation before usually use a “nearest neighbors” approach, basically for simplicity’s sake. But not all “spillover phenomena” are necessarily primarily based on geographic proximity. In COVID’s case, the car-based trips may account for much more epidemiological interaction than local pedestrian trips. So one of the ways to think about Safegraph data is as a “mobility-weighted neighbor matrix” that can be used to replace the “nearest neighbor matrix” in spatial autocorrelation. And the idea of creating contiguous clusters would be similar, if you were doing something like a stepped wedge RCT and wanted to ex ante define groups that minimized interference.

Derek, you should reach out to Filipi Nascamiento Silva at IU–he has done some great work on spatial clustering using the SafeGraph SDM data.