mstrimas
commented
4 years ago I found this whole episode a challenge to work with and I'll probably skip or greatly simplify when teaching. ggplot2 legends don't do well when plotting multiple vector layers with different symbol types. Having to detour to explain the needed overrides in the challenge solution seems like a huge time sink that will almost certainly lead to confusion. Glossing over the legend issue, essentially what's going on in the lesson by not using show.legend = "points", makes things simpler but doesn't prepare learners for the inevitable challenges they'll face when working with their own data.
Unfortunately, I don't have a good solution here, the reality is ggplot2 has limitations when creating anything but the simplest maps of spatial data. It would be great to have the time to delve into how to deal with this (e.g. overrides), but there simply isn't enough time in an already packed schedule. The other obvious solution is a different package that offers more flexibility for plotting spatial data, but then you have to spend time teaching that new tool, which doesn't seem feasible.
Anyhow, maybe this is worth some discussion here...
As per https://github.com/datacarpentry/r-raster-vector-geospatial/pull/232