Making a Map of Exeter using geojson and ggplot2

A few weeks ago I discovered the Exeter Data Mill, an amazing project where really interesting data about Exeter are collected and published. This included a very interesting series of files called “Exeter shapes and locations” available under the Open Government License.

I’m not usually one for maps, but this struck me as a good opportunity to learn a new skill. I decided to try to make a map of Exeter postcode districts. Then, because a map on its own is not very helpful, I wanted to make up some data I could use to make a rudimentary heatmap on top of the postcode areas.

The first step was loading all of the packages I would need for this. After much trial and error I decided that I needed geojsonio, ggplot2, and broom.


I found nearly all the info I needed to make this map by using the R graph gallery.

I then needed to download the geojson data and save it as an R object.

spdf <- geojson_read("map/EX_Sectors.geojson", what = "sp")

I then needed to tidy the data to get it into a usable format using the tidy() function of the broom package.

spdf_fortified <- tidy(spdf, region = "name")

The next step is making a lookup to identify different regions. I got this idea from the ggplot2 website.

I want to make sure I have every postcode region as a single item, so I wanted to delete the duplicates.

spdf_fortified$postcode <- gsub(" .*", "", spdf_fortified$group)
values <- data.frame(postcode = unique(spdf_fortified$postcode),
                     value = runif(length(unique(spdf_fortified$postcode)),0,0.99))

##   postcode     value
## 1      EX1 0.1125664
## 2     EX10 0.6160764
## 3     EX11 0.6031820
## 4     EX12 0.6171456
## 5     EX13 0.8523062
## 6     EX14 0.6339075

The values dataframe is a dataframe with 2 columns. The postcode area and the value, which is our imagined data. Perhaps it is the proportion of the population with a degree or something to that effect.

Next we need to merge the two dataframes, matching based on postcode.

merged_df <- merge(values,spdf_fortified, by = c("postcode"))

And finally plot the results. geom_polygon does nearly all of the work for us, we only need to add the alpha value to make sure the different values appear at different levels of opacity.

ggplot() +
  geom_polygon(data = merged_df, aes( x = long, y = lat, group = group, alpha = value), fill="#489D5D", color="white") +
  theme_void() +

Dr Greg Stride
Dr Greg Stride

My research interests include UK elections, election administration and public opinion