In jr-packages/jrSpatial: Jumping Rivers: Spatial Data Analysis using R

Start by loading the correct libraries

library("tmap")
library("jrSpatial")

Question 1 - first steps in tmap

We're going to be using the same UK shape data from the first practical but don't worry, we'll not get you to read it in again. We've got it stored inside the course package

data("ukgeom", package = "jrSpatial")

a) Using tmap, plot the 41 regions within uk_shape

tm_shape(ukgeom) +
  tm_fill() +
  tm_borders()

b) Colour the polygons by the population in 2017

tm_shape(ukgeom) + 
  tm_fill(col = "pop_2017") +
  tm_borders()

c) The legend isn't a very good position, change it to either the top left or outside of the graph. Hint: If you don't like the position of the legend outside the graph, using legend.outside.position to change it.

tm_shape(ukgeom) +
  tm_fill(col = "pop_2017") +
  tm_borders() +
  tm_layout(legend.position = c("left", "top"))
# or
# tm_layout(legend.outside = TRUE)

d) By default the legend title is the variable name, we can change this by using the title argument within tm_fill(). Have a go!

tm_shape(ukgeom) +
  tm_fill(col = "pop_2017",
          title = "UK population 2017") +
  tm_borders() +
  tm_layout(legend.position = c("left", "top"))

e) Have a go at changing the colour palette. See http://colorbrewer2.org for some options!

tm_shape(ukgeom) +
  tm_fill(col = c("pop_2015", "pop_2015"),
          palette = "PuRd") +
  tm_borders()

f) Colour the polygons by both the population of 2015 and 2017. Can you see any major differences?

tm_shape(ukgeom) +
  tm_fill(col = c("pop_2015", "pop_2017")) +
  tm_borders()
# no major difference

Question 2 - Overcrowding

Let's say we're interested in evaluating overcrowding in the UK. Absolute population isn't a very good measure because each region has a different area in total. A better measure would be area per person. We can work this out as we have an area column in the shape data. Below is the code to create the column then plot the data

library("dplyr")
ukgeom = mutate(ukgeom, areapps = area / pop_2017 * 1000)
# * 1000 to change units from km^2 to m^2
tm_shape(ukgeom) + 
  tm_layout(legend.position = c("left", "top")) +
  tm_fill("areapps",
          title = "Area per person") +
  tm_borders()

Because the area at the top of Scotland is so vast an unpopulated compared to the rest of England, we're getting a very skewed colour scale. Change the breaks within tm_fill() to something more appropriate

tm_shape(ukgeom) +
  tm_layout(legend.position = c("left", "top")) +
  tm_fill("areapps", 
          title = "Area per person",
          breaks = c(seq(0, 20, 4), 100)) +
  tm_borders()

Question 3 - Hills!

As well as our regional polygons data for the UK, we also have a data set of points for the highest hill (or mountain) in each region. You can load the data like so

data("ukhills_big", package = "jrSpatial")

We can add several layers of different shapes onto a tmap, just by adding another tm_shape() like so

tm_shape(ukgeom) + 
  tm_fill() + 
  tm_borders() + 
tm_shape(ukhills_big) + 
  tm_bubbles()

a) Try changing the colour of the mountain points depending on it's height, and try changing the size of the point using the size argument.

tm_shape(ukgeom) + 
  tm_fill() + 
  tm_borders() + 
tm_shape(ukhills_big) + 
  tm_bubbles(col = "Metres", 
             size = 0.5)

b) Run the following line of code and then re-run the code for your graph in question 3a). You should get an interactive version of the map!

tmap_mode("view")

c) Currently the text that appear when you hover over a region is the region code, instead of the region name. For mountains this text is the region, instead of the mountain name. Try adding the arguments id = "region" to tm_fill() and id = "Mountain" to tm_bubbles(). Try clicking on the shapes!

tm_shape(ukgeom) + 
  tm_fill(id = "region") + 
  tm_borders() + 
tm_shape(ukhills_big) + 
  tm_bubbles(col = "Metres", 
             size = 0.1, 
             id = "Mountain")