Hello,

I came across this correlation plot type on Sthda and thought of applying it to Google Analytics data. Here’s the article from the package creator: http://www.sthda.com/english/wiki/ggcorrplot-visualization-of-a-correlation-matrix-using-ggplot2

If you’d like to apply this to Google Analytics in R, there are a few steps you can follow [Full script is at the end]:

• Use the GCP service credentials key for your R project

• Make the API call to GA via googleAnalyticsR. In this example, I pulled sessions, bounceRate, pageviewsPerSession, users, goal1Completions as metrics along with date and dayOfWeek as my dimensions. Once you pull your data, it should look like this.

• Install and load ggcorr package

• Run the correlation function in base R via… corr <- round(cor(data[ , c(3:7)]),1) . As you need numeric columns for correlation, I’ve asked R to use columns 3:7 only. If you try running this with non numeric columns, you’ll get an error in console “row x needs to be numeric”

• Last bit is to just apply the ggcorrplot command to your corr object and put labels = TRUE to show correlation in the plot

That’s about it.

Correlations can vary between -1 and 1. Sessions and Users have a correlation of 1. PagesPerSession and BounceRate have a negative correlation of -0.5, makes sense as they should move in the opposite direciton. The STHDA article has more options on visualizations. You may also run ?ggcorrplot() in R to read the documentation.