如何在散点图中圈出一组点并获取其行？

Question

I generated some plots with ggplot and colored them by the magnitude of variable x.

I would like to know the identity of the cells that form the main purple cloud in the plot seen below.

Is there a way to encircle this area and return a list of rows from the dataframe that are in the circle?

Answer 1

In principle one could use ggmap::gglocator() to select a bunch of points, construct a polygon, then use spatial tools to find points in the polygon, but it would take a lot of work to put those pieces together. If it were me I would (1) select all points with x>(-15) & y>0 (2) from that set, select all points where y>40-(4/3)*x (eyeballing the intercept slope of a line that forms the lower left boundary of the cluster). (In fact, step #2 would probably work by itself without running step #1 first.)

With a little bit more work one could write a tool to generalize this that would define bounding lines from pairs of gglocator() points ...

Answer 2

I realize this is an old post - but I've found a nice bit of code here to deal with the first part of the question (see "Scatterplot With Encircling"):

http://r-statistics.co/Top50-Ggplot2-Visualizations-MasterList-R-Code.html

UPDATE: Link context - someone wiser and more experienced than me has documented that there is a function (geom_encircle) which provides a solution to the first part of the question. This function integrates nicely into ggplot2 syntax. I just happened to notice both this SO question and the linked website at the same time, and thought it would be useful for future reference for folks looking for the same thing (and didn't want to make it seem like i should get credit). geom_encircle requires the ggalt package.

The full link to take you directly to the relevant place on the page: http://r-statistics.co/Top50-Ggplot2-Visualizations-MasterList-R-Code.html#Scatterplot%20With%20Encircling