Error Bars

[oldoc63]

You've learned to represent data as bars of different heights. Sometimes, we need to visually communicate some sort of uncertainty in the heights of those bars. Here are some examples:

• The average number of students in a 3rd grade classroom is 30, but some classes have as few as 18 and others have as many as 35 students.
• We measure that the weight of a certain fruit was 35g, but we know that our scale isn't very precise, so the true weight of the fruit might be as much as 40g or as little as 30g.
• The average price of a soda is $1.00, but we also want to communicate that the standard deviation is $0.20.

To display error visually in a bar chart, we often use error bars to show where each bar could be, taking errors into account.

[oldoc63]

Each of the black lines is called an error bar. The taller the bar is, the more uncertain we are about the height of the blue bar. The horizontal lines at the top and bottom are called caps. They make it easier to read the error bars.

If we wanted to show an error of +/- 2, we would add the keyword yerr=2 to our plt.bar command. To make the caps wide an easy to read, we would add the keyword capsize=10:

[oldoc63]

Like the list of x-axis labels, Matplotlib reads this in the same order as the list of y-values. So, the first index of your error list should correspond to the first index of your y-values list, and so on.

[oldoc63]

1. For someone who is learning about the different drink types at MatplotSip, a bar chart of milk amounts in each drink may be useful. We have provided the ounces_of_milk list, which contains the amount of milk in each 12oz drink in the drinks list. Plot this information as a bar chart.

[oldoc63]

2. According to different barista styles and measurement errors, there might be variation on how much milk actually goes into each drink. We've included a list error, with and error of 10% on each amount of milk. Display this error as error bars on the bar graph.

[oldoc63]

3. Add caps of size 5 to your error bars.