10 Chart: Bar Chart

10.1 Overview

This section covers how to make bar charts

10.3 Simple examples

My eyes were bigger than my stomach. Much simpler please!

Let’s use the HairEyeColor dataset. To start, we will just look at the different categories of hair color among females:

## # A tibble: 4 x 2
##   Hair  Total
##   <fct> <dbl>
## 1 Black    52
## 2 Brown   143
## 3 Red      37
## 4 Blond    81

Now let’s make some graphs with this data.

10.3.1 Bar graph using base R

We recommend using Base R only for simple bar graphs for yourself. Like all of Base R, it is simple to setup. Note: Base R expects a vector or matrix, hence the double brackets in the barplot call (gets columns as lists).

10.3.2 Bar graph using ggplot2

Bar plots are very easy in ggplot2. You pass in a dataframe and let it know which parts you want to map to different aesthetics. Note: In this case, we have a table of values and want to plot them as explicit bar heights. Because of this, we specify the y aesthetic as the Total column, but we also have to specify stat = "identity" in geom_bar() so it knows to plot them correctly. Often you will have datasets where each row is one observation and you want to group them into bars. In that case, the y aesthetic and stat = "identity" do not have to be specified.

10.4 Theory

  • For more info about plotting categorical data, check out Chapter 4 of the textbook.

10.5 When to use

Bar Charts are best for categorical data. Often you will have a collection of factors that you want to split into different groups.

10.6 Considerations

10.6.1 Not for continuous data

If you are finding that your bar graphs aren’t looking right, make sure your data is categorical and not continuous. If you want to plot continuous data using bars, that is what histograms are for!

10.7 Modifications

These modifications assume you are using ggplot2.

10.7.2 Reorder the bars

With both base R and ggplot2 bars are drawn in alphabetical order for character data and in the order of factor levels for factor data. However, since the default order of levels for factor data is alphabetical, the bars will be alphabetical in both cases. Please see this tutorial for a detailed explanation on how bars should be ordered in a bar chart, and how the forcats package can help you accomplish the reordering.

10.7.3 Facet Wrap

You can split the graph into small multiples using facet_wrap() (don’t forget the tilde, ~):

10.8 External resources







with