R visualization of the correlation of a point distribution to the attributes of the polygons they are located in

Question

Situation:
I have two spatial data frames, one of type polygon (areas with different average wind speed levels) and one of type point (wind mills). I read the polygon shapefile with readOGR, the points were imported as csv and I added coordinates and a projection to it.
What I have:
I use the function over() from the sp-package in combination with summary to determine how many wind mills are placed in areas of each category of average wind speed. Afterwards, I can compare this distribution to the general distribution of wind speed levels and get some kind of result.
Question:
I want a plot that visualizes the relationship between the positions of the wind mills and the wind speed levels. (So hopefully see that wind mills are preferably placed in polygons with high wind speed.) How can I accomplish this?

library('raster')
library('rgeos')
library('sp')
library('rgdal')

polygons = readOGR(dsn = "./North_South_Dakota_Wind_High_Resolution", layer = "ndsd_50mwind")
wind_mills = read.csv('NDGISHubData/NDHUB_WINDTURBINES.csv')
# Assignment modified according
coordinates(wind_mills) <- ~ LONGITUDE + LATITUDE
# Set the projection of the SpatialPointsDataFrame using the projection of the shapefile
proj4string(wind_mills) <- proj4string(polygons)
# get wind mill distribution
windpower_values = over(wind_mills, polygons)
# remove NAs
windpower_distribution = lapply(windpower_values, function(x) x[!is.na(x)])
# create summaries for correlation comparison
# WPC is the attribute containing the wind speed level
a = summary(windpower_values$WPC)
b = summary(polygons@data$WPC)
# ToDo: spatial analysis?

Answer 1

This sounds like something you could do with a scatter plot. You have average wind speeds by polygon and the number of mills per polygon. I would recommend you convert those spatial dataframes to simple features objects in order to simplify how to do this, then you can use ggplot or any other plotting package to make a variety of plots.

If you're thinking more cartographically, then there a couple of options also. If you want to be fancier there is a bivariate choropleth approach where you could potentially locate areas of windmill concentration and highest average windspeed compared to other combinations. Or you could do something like create a cut off value for windspeed and windmill concentration and only highlight those polygons that meet the cutoff values.