Animated Spatial Clustering

This page has four animated choropleth maps that show the spatial clustering in our variables for each week in 2020. Counties classified as hotspots are colored red and counties classified as coldspots are colored blue. For more information on how counties are classified as hotspots or coldspots see the About Page.

The top row of plots includes data from Cuebiq while the bottom row is from Safegraph. The column on the left shows the sources' measures of mobility while the column on the right is of measures of sheltering. If all of these variables are showing the same phenomenon then we would expect the same patterns in the each column and inverted colors across the rows. However, one of the points to be made by comparing different variables and different sources is that they don't show the exact same patterns which could lead to different predictions in models using different data.

Click on a state above to see a table of the state's county values.

One of the first things to notice is how sparse the classifications are in Safegraph's mobility measure in the column on the left. If you turn on the cumulative map option and play the animation to the end, we can see all of the counties that were once classified in the year. While the other three maps are mostly colored in, this map has more counties that were never significant.

On the 13th week, 3/23-3/29, we can see an abrupt change in the maps. These dates correspond to just after the declaration of a national emergency which is listed on the timeline events. In all of the weeks leading up to these dates, we see a large number of counties classified as coldspots in Cuebiq mobility and hotspots in the sheltered column, particularly in the west, which go away as the pandemic begins. However, particularly in the South around Mississippi and Alabama, we don't see this change as the classification of these counties remain constant throughout the year.