When working with geospatial data, memory consumption and computation time can become quite a problem, since these datasets are often very large. You may have very granular, high resolution data although that isn’t really necessary for your use-case, for example when plotting large scale maps or when applying calculations at a spatial level for which lower granularity is sufficient. In such scenarios, it’s best to first *simplify* the *geospatial features* (the sets of points, lines and polygons that represent geographic entities) in your data. By simplify I mean that we generally reduce the amount of information that is used to represent these features, i.e. we remove complete features (e.g. small islands), we join features (e.g. we combine several overlapping or adjacent features) or we reduce the complexity of features (e.g. we remove vertices or holes in a feature). Since applying these operations comes with information loss you should be very careful about how much you simplify and if this in some way biases you results.

In R, we can apply this sort of simplification with a few functions from the packages sf and, for some special cases explained below, rmapshaper. In the following, I will show how to apply them and what the effects of their parameters are. The data and source code are available on GitHub.

Read More →