Visualizing statistical data more intuitively

Science is based on statistics. We propose hypotheses from observations. We test and reject the null hypothesis when its probability, the p value, is lower than a threshold, so that observed phenomena are not likely arisen from mere chance; that is, our proposed hypothesis is supported.

There are some specific plot types that can ease the visualization of descriptive and inferential statistics. We will first revisit more variants of bar charts–stacked bar chart and layered histograms, which are commonly used in scientific publications to summarize and describe data.

Stacked bar chart and layered histogram

In Chapter 4, Visualizing Online Data, we showed the procedures to create bar charts using Matplotlib and Seaborn. It is little known that the pandas package can be used for visualization, as most people only concentrate on its data analysis capabilities. Since the pandas visualization module was built on top of Matplotlib, we can exploit the powerful...