The world’s leading publication for data science, AI, and ML professionals.

DIY: Make Your Own Scatter Plot

To use the web application: click here

scatter plot- R application (Image by author)
Scatter Plot Web application screen grab (Image by Author)

_To use the web application: click here_

Creating sensible scatter plots is a huge challenge, thus I’ve created, ground up, a web application that allows you to create good scatter plots. So use your own data, configure your plot to your liking and save the plot for further use in presentations/ reports or elsewhere.

Note: Neither your input data, nor the plot persist after the application is terminated.

About The Web Application

The base plot has been created using R and ggplot2 visualisation library. This was converted into a web application using R shiny and the application is hosted on shinyapps.io

Challenges with scatter plots

Here are some of the key challenges with making meaningful scatter plots:

  • Identifying patterns in the (many) data points
  • Drawing attention to specific areas of the plot
  • Highlighting some data points
  • Understanding relationship among the 2 variables

Improving your scatter plot

Here are some suggestions for making better scatter plots:

  • Highlight data points of interest e.g. top/ bottom 5%, middle 80%
  • Use anchors/ cut-offs independently for X & Y variables e.g. mean, median, quartiles, extremes (e.g. bottom/ top 10%)
  • Fit a curve (i.e. model) to understand relationship among the 2 variables
  • Fit a curve (i.e. model) to study departures from the modelled (i.e. average) line

Happy (scatter) plotting!

To learn more about principles to create meaningful plots, do read my Medium article ‘Creating Good Meaningful Plots: Some Principles‘.


Related Articles