Showing Distributions

Distribution charts help us understand the underlying shape of a dataset: where values are concentrated, the presence of outliers, and the overall variance.

Histograms

A histogram looks like a bar chart, but it groups continuous data into bins. The area of the bar represents the frequency.

Crucial Rule: The choice of bin size drastically alters the shape of the histogram. Too few bins (oversmoothing) hide important details. Too many bins (undersmoothing) create noise.

Box Plots (Box-and-Whisker)

Excellent for comparing distributions across multiple categories simultaneously without taking up much space. It encodes the five-number summary: minimum, first quartile, median, third quartile, and maximum.

Common Mistakes

• Using a Bar Chart for Averages: Showing an average with a bar chart hides the underlying variance. A bar implies data exists all the way down to zero. Use a point with error bars or a box plot instead.

Next Steps

If you want to see if two distributions relate to each other, see Relationships.