Simple data mining and plotting data on a map with ggplot2

April 02, 2012


A Facebook group for psychologists in Sweden created a document where each member could type in were they live and what they work with. At the moment 192 psychologists have added their information to the document. This made me think about how to plot this information using R, and especially how to read data in this format. I was also interested in how to plot this information geographically on a map of Sweden and represent the number of individuals by the size of a circle over the corresponding city.

The data

The document with the data was ordered like this:


Parsing the data

To get the data into R I copied the Facebook document to a txt-file and used scan() to read it. This saves the text into a vector with each line representing one row. I wrote a function to parse the text and to create a data.frame with the number of people of each city.

Creating the map

To get a map of Sweden I used the OpenStreetMap-package. But since I wanted to plot data on this map I had to grab the coordinates for each city. This was accomplished by using data from the “world.cities”-set which is included in R. However, 5 cities weren’t included in this set so I looked up their coordinates on Google Maps and added them manually. Lastly I converted latitude and longitude into Mercator coordinates since this is the format used when plotting from the OpenStreetMaps-package.

Creating the plot

I used autoplot() to plot the map with ggplot2(). The city data was added as geom_point()-layer with size = number of psychologists in the city.

The resulting plot

where do we live. By Kristoffer Magnusson Figure 1. Plot of where people live, with data taken from a Facebook document.

The R Code

Written by Kristoffer Magnusson a researcher in clinical psychology. You should follow him on Twitter and come hang out on the open science discord Git Gud Science.


Published April 02, 2012 (View on GitHub)