Is geom_map obselete in the sf world?

ggplot2

#1

A major advantage of ggplot2::geom_map seemed to be smoothing the many:1 impedance mismatch between fortifyed spatial objects (points as row item) and non-geometric attributes (meaningful entities as the row level).

Since the sf package now resolves that directly, is geom_map basically dead?


#2

I'd say it's entering old age, providing support for sp work that hasn't yet been ported. Recently, I saw Hadley talking about the demise of fortify in favor of broom. And, of course, sp itself is being replaced by sf by its own developers.

It took me a bit to grok sf, but being able to get a geometry object into a tibble is huge. Points, lines and polygons of various stripes al extractable. I've run across a couple of pain points, one of which is that extracting centroids is a two step process that yields a matrix that has to be converted and inserted based on blind faith (so-far) of a 1:1 correspondence with the associated labels. I've had good luck with thematic mapping state abbreviations.