In this practical, you will visualize the development of wealth in Basel across time.

0 - Preliminaries

  1. If you haven’t done so already, load the tidyverse, patchwork, and the taxation.csv data.
library(tidyverse)
library(patchwork)
basel <- read_csv('1_data/taxation.csv')

1 - Summarize

  1. Begin by aggregating the mean, median, and Gini of wealth across the different quarters to obtain average estimates for each individual year. Save the result as avg_wealth.
avg_wealth <- basel %>% 
  group_by(year) %>% 
  summarize(avg_wealth_mean = mean(wealth_mean),
            avg_wealth_median = mean(wealth_median),
            avg_wealth_gini = mean(wealth_gini))

2 - Plot

  1. Start a ggplot using ggplot() by piping the avg_wealth data set into the function’s data argument.
avg_wealth %>% 
  ggplot()

  1. Running the previous chunk should have created a blank canvas in your bottom right RStudio panel. Now set up the structure of the plot using the mapping argument and the aes() helper function. Place year on the x-axis and avg_wealth_mean on the y-axis.
avg_wealth %>% 
  ggplot(mapping = aes(x = year, 
                       y = avg_wealth_mean))

  1. The updated plot should now contain axis labels, tick marks and labels, and grid lines. It’s time to add some geometric shapes to show the data. Add geom_point().
avg_wealth %>% 
  ggplot(mapping = aes(x = year, 
                       y = avg_wealth_mean)) +
  geom_point()

  1. The figure should show 17 points describing a clear upward trend: Baselers have become quit a bit more wealthy. Would an additional geom_line() help emphasize the trend?
avg_wealth %>% 
  ggplot(mapping = aes(x = year, 
                       y = avg_wealth_mean)) +
  geom_point() + geom_line()

  1. Maybe geom_smooth() would be an even better choice? Try adding it instead of geom_line(). You can set se = FALSE inside geom_smooth(), to omit the confidence interval.
avg_wealth %>% 
  ggplot(mapping = aes(x = year, 
                       y = avg_wealth_mean)) +
  geom_point() + geom_smooth(se = FALSE)

3 - Adding a variable

  1. To illustrate the development of the medians in the same plot, add geom_point() and either geom_line() or geom_smooth() for avg_wealth_median. Inside these additional geoms specify that they should show avg_wealth_median on the y-axis using the aes(). Note that for geoms the first argument usually is mapping, which is why we can omit its name.
avg_wealth %>% 
  ggplot(mapping = aes(x = year, 
                       y = avg_wealth_mean)) +
  geom_point() + geom_line() + 
  geom_point(aes(y = avg_wealth_median)) + 
    geom_line(aes(y = avg_wealth_median))

  1. Looks like not every Baseler got a lot richer. Increases in medians are much more moderate and there is quite a gap between means and medians. What does that imply for the development of the Gini?

4 - Facets

  1. Let’s try to bring the Gini into the picture. Since the Gini is on a completely different scale, the best way to do this is facets. Before we can go there, however, you need to transpose avg_wealth to the long format. Let the new variables be called "variable" and "value". Save the result as avg_wealth_long.
avg_wealth_long <- avg_wealth %>% 
  pivot_longer(c(avg_wealth_mean, avg_wealth_median, avg_wealth_gini),
               names_to = "variable",
               values_to = "value")
  1. Use avg_wealth_long and facet_wrap() to create a plot that illustrates the development for all three variables with points and lines. To do this, value must go on the y-axis and variable must be used as the faceting variable.
avg_wealth_long %>% 
  ggplot(mapping = aes(x = year, 
                       y = value)) +
  geom_point() + geom_line() +
  facet_wrap(~ variable)

  1. Not ideal, right? The problem is that, per default, ggplot tries to use the same scales for the different facets. Set scales = "free" inside the facet_wrap function. Depending on your screen layout you may also want to set ncol = 1 to have the facets sit on top of each other.
avg_wealth_long %>% 
  ggplot(mapping = aes(x = year, 
                       y = value)) +
  geom_point() + geom_line() +
  facet_wrap(~ variable, scales = "free", ncol = 1)

  1. For Baselers everything has been going up: The mean wealth, the median wealth, and the Gini. Development in the Gini seems mild, however, note that it sits almost at the maximum of the scale. The Gini can maximally be 1 and that only if one person has all of the wealth and everyone else nothing. Part of the reason for the mild development, thus, likely is that extreme changes in the distribution of wealth are necessary for the Gini rise by noticeable amounts.

5 - Multiple plots

  1. Ok, let’s try to compose the plot of the three wealth metrics in a way that gives the Gini more attention. Specifically, use the patchwork package to create a plot with rows where the development of the Gini takes up the entire upper row, while the other two metrics share the lower row. To do this, first create three different plots, one for each of the three panels.
gini_plot <- avg_wealth_long %>% 
  filter(variable == "avg_wealth_gini") %>% 
  ggplot(mapping = aes(x = year, 
                       y = value)) +
  geom_point() + geom_line() 

mean_plot <- avg_wealth_long %>% 
  filter(variable == "avg_wealth_mean") %>% 
  ggplot(mapping = aes(x = year, 
                       y = value)) +
  geom_point() + geom_line() 

median_plot <- avg_wealth_long %>% 
  filter(variable == "avg_wealth_median") %>% 
  ggplot(mapping = aes(x = year, 
                       y = value)) +
  geom_point() + geom_line()
  1. Now, use patchwork syntax to compose the plot.
gini_plot / (mean_plot + median_plot)

  1. Finally, add plot_annotation(tag_levels = "A") so that the labels are lettered according to the alphabet.
gini_plot / (mean_plot + median_plot) + plot_annotation(tag_levels = "A")

6 - Project work

  1. Try to use what you have learned in this section to visually explore your project data and possibly create a basic version of your data visualization.