Circles in sf

[roaldarbol]

Hi there! I'm just beginning to get a hold of what is possible with sf, but am super excited so far!

One thing I am trying to figure out what is possible with sf in relation to circles. So a few questions (and I'm happy to create issues and/or PRs where it would make sense afterwards) in case anyone has any knowledge and would like to chat about it:

• Is it possible to fit a circle to a set of points (which do not themselves make a perfect circle), similar to st_ bbox? I'm imagining three possible methods/modes would be possible: inside, which makes the largest circle that fit inside all the points; outside which is the smallest circle that fit outside all of the points; mean which minimises the mean of all the distances (as in https://r-spatial.github.io/sf/reference/st_nearest_feature.html). outside would be the same as st_bbox.
• Making zones should be relatively easy I think, with st_ buffer from the central point.
  • Maybe there would be a way to divide a circle up into "slices (of pie)" as an alternative set of zones?

Thanks, and looking forward to feedback!

[edzer]

The function is there in GEOS, GEOSMinimumBoundingCircle_r, but we don't have it interfaced in sf. Implementing it in sf is reasonably trivial, but involves conditioning on the GEOS version in which it appeared.

[roaldarbol]

Awesome! Seems to have been implemented in 3.9.1 according to the docs of the R package of the same name (in which it was implemented in 0.1.0) https://paleolimbot.github.io/geos/news/index.html.

[roaldarbol]

@edzer Should I open an issue for this to keep it on the radar?

[edzer]

That's OK.