Tuning your bar and point charts

Visualization Best Practices in R

Nick Strayer

Instructor

A busy bar chart

who_disease %>% 
  filter(region == 'EMR', disease == 'measles', year == 2015) %>% 
  ggplot(aes(x = country, y = cases)) +
    geom_col()

Flipping the bar

geom_bar() and geom_col() don't allow categories on y-axis

busy_bars <- who_disease %>% 
filter(region == 'EMR', disease == 'measles', year == 2015) %>% 
ggplot(aes(x = country, y = cases)) +
geom_col()

So we have to flip!

busy_bars + coord_flip() # swap x and y axes!

Excess grid

No need for parallel grid lines in bars
In point charts, only grids in line with point locations are needed

Removing vertical grid

plot <- who_disease %>% 
  filter(country == "India", year == 1980) %>% 
  ggplot(aes(x = disease, y = cases)) +
    geom_col()

# Remove vertical grid lines
 plot + theme(
    panel.grid.major.x = element_blank()
  )

Lighter background for point charts

Default gray background can be too low-contrast for points
theme_minimal() is a quick fix
Making points bigger helps too

who_subset %>% 
  ggplot(aes(y = reorder(country, cases_2016), x = log10(cases_2016))) +
  # Point size increased
  geom_point(size = 2) +
  # Theme minimal for light background
  theme_minimal()

Let's try it out!

Visualization Best Practices in R