They're indexing the popularity of e-readers in different MSAs by looking at how many are for sale on secondary markets, as a % of what's for sale on the secondary markets in that city. This seems like a really poor proxy for how popular e-readers are in those locations.
There are several things they're not correcting for:
* The more someone uses their e-reader, the less likely it is to end up on a secondary market - either because you're using it, or you still intend to use it, or you broke it. (This is what happened to my last Kindle - I used it heavily, but eventually something in my bag rammed into the screen and broke it to a point it was unsaleable.)
* Some cities may have a larger volume of items per-capita on a secondary market because people have more to sell. E.g. if you have a larger house with a garage, you probably have more items to potentially resell than someone with a tiny apartment.
* Cities with a lower proportion of smartphone / tablet users might buy more kindles. That doesn't mean they're consuming fewer ebooks.