A “Joy plot” style map of the Isle of Wight

This is a relatively shameless copy of the wonderful Helen McKenzie tutorial on how to produce “joy plot” style maps, which is in turn a simplified version of Travis White’s tutorial. I’ve included a little more detail on some of the finer points of the process in areas that tripped me up a bit. I’m going to apply this style to my beloved Isle of Wight, an island just off the south coast of England. We’ll be using a combination of QGIS and R to create the map, with Adobe Illustrator to add the finishing touches.

Elevation data

There are a huge number of sources for this, if you’re considering a small area it might be worth looking for some LiDAR data, but that resolution (1-2 m) isn’t necessary for this. I went with SRTM30m.

• https://earthexplorer.usgs.gov
• https://apps.sentinel-hub.com/eo-browser
• https://dwtkns.com/srtm30m/
• UK LiDAR: https://environment.data.gov.uk/DefraDataDownload/?Mode=survey

QGIS

• Import into QGIS.
  • We need to set the CRS of both our project and our raster layer. I like to use OSGB 1936 (EPSG:7405) for the UK as it gives less of a squished (technical term) perspective in comparison to the commonly used WGS 84 (EPSG:4326).
  • Set the project CRS in Project>Properties…>CRS
  • QGIS will convert a layer CRS to the project CRS on-the-fly if we double-click the layer and assign a CRS in “Source”, but we need to permanently reproject the layer.
  • Use the Warp (reproject) tool from the processing toolbox to set a target CRS of OSGB 1936 (EPSG:7405) and run.
  • Now you have the layer reprojected the units should be given in meters, which will be useful when creating a grid.
• Create a polygon and crop the raster to that extent.
  • Create a new shapefile layer and draw a polygon (Add polygon feature). You don’t have to be super precise with this as we can set the minimum height at which a line is drawn in R. You can check the elevation at your mouse pointer with the value tool.
  • Use Raster>Extraction>Clip Raster by mask layer… to crop your raster down to the shape of the polygon. Make sure to finish editing the polygon before trying to clip the raster or the process will throw an error.
• Create a grid and assign elevation data to lines.
  • Create a grid using Vector>Research tools>Create grid. Try spacings of around 200 m.
  • Clip the lines to the raster layer (Vector>Geoprocessing Tools>Clip).
  • Delete the vertical lines.
  • Run the tool “Generate Points (Pixel Centroids) Along Line” from the processing toolbox.
  • Install the point sampling tool. Plugins > Manage and Install Plugins… > search for “Point Sampling Tool”. Go to Plugins > Analyses and open the tool, now select “Points along lines: id (source point)” and your raster layer by ctrl+click. Select where to save the output and run the tool.
  • Assign elevation data from the raster to each of the points using “Add X/Y fields to layer” from the processing toolbox.
  • Export as .csv by right-clicking the layer and using Export>Save As…

RStudio

You’ll need to install the packages ggplot2, ggridges, and dplyr for this. Play around with the scale value to exaggerate the elevation. Set the rel_min_height based on the height of your water level, so that only the land is drawn.

library(ggplot2)
library(ggridges)
library(dplyr)

transects <- read.csv("data.csv")

head(transects)

names(transects)[1]<-"ID"
names(transects)[3]<-"Elevation"
names(transects)[4]<-"Lon"
names(transects)[5]<-"Lat"

joy <- ggplot(transects,
              aes(x = Lon, y = Lat, group =Lat, height = Elevation))+
  geom_density_ridges(stat = "identity",
                      scale = 25,
                      fill = "black",
                      color = "white",
                      rel_min_height = 0.005)+
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        panel.background = element_rect(fill="black"),
        axis.line = element_blank(),
        axis.text.x = element_blank(),
        plot.background = element_rect("black"),
        axis.title.x = element_blank(),
        axis.text.y = element_blank(),
        axis.title.y = element_blank())
joy

ggsave("joy_iow.pdf", width = 1920, height = 1080, units = "px")

Final touches

Import into your choice of image editor to add a border or a title. I went with Trajan Pro for the font, as used on Joy Division’s “Closer” cover. I think it’s a bit more interesting than Helvetica, as used on “Unknown Pleasures”.