Plotting Pie Chart

Pie Chart: A Part-of-the-Whole Visualisation

A pie chart is a circular graph divided into sectors, where each sector represents a proportion of the total. The entire circle stands for the complete dataset, and each slice shows how much of that whole a particular category contributes. This makes pie charts ideal for displaying numerical data proportionally — for example, showing how forest cover is distributed across different states.

In Matplotlib, you create a pie chart from a DataFrame using df.plot(kind='pie'). There is an important rule: you must either specify a column using the y parameter or set subplots=True. If you pass no column reference and set subplots=True, a separate pie chart is drawn for every numerical column in the DataFrame.

How Labels Work in a Pie Chart

When you plot a pie chart from a DataFrame, the default labels come from the DataFrame's index. In the textbook's first example (Program 4.16), a DataFrame holds planet names as the index and their mass and radius as columns. Passing 'mass' to the plot() function produces a pie chart where each slice is labelled with the planet's name — because those names are the index values.

A Worked Example: Forest Cover Data

Consider a dataset of seven north-eastern Indian states. The DataFrame has two columns: GeoArea (total geographical area in sq km) and ForestCover (forested area in sq km). The state names serve as the index.

To plot a pie chart showing only the forest cover, you write:

df.plot(kind='pie', y='ForestCover', title='Forest cover of North Eastern states', legend=False)
plt.show()

Notice the parameter legend=False. This removes the legend from the output. Without it, Matplotlib would display a legend alongside the pie chart, which is often redundant because the slice labels already identify each category. The resulting chart (Figure 4.21) shows seven slices, each labelled with a state name and sized proportionally to that state's forest cover.

Customising the Pie Chart

The textbook introduces two powerful customisation options: explode and autopct.

Explode lets you pull one or more slices away from the centre of the pie, drawing attention to them. You provide a list of values, one per slice, where each value is the fraction of the radius by which that slice should be separated. A value of 0 means the slice stays in place; 0.1 means it is shifted outward by 10% of the pie's radius.

Autopct displays the percentage value of each slice directly on the chart. You pass a format string — for example, "%.2f" shows percentages with two decimal places.

Here is the customised version from Program 4.18:

exp = [0.1, 0, 0, 0, 0.2, 0, 0]   # explode first slice by 0.1, fifth by 0.2
c = ['r', 'g', 'm', 'c', 'brown', 'pink', 'purple']

df.plot(kind='pie', y='ForestCover', title='Forest cover of North Eastern states',
        legend=False, explode=exp, autopct="%.2f", colors=c)
plt.show()

The colors parameter lets you assign a specific colour to each slice. In this example, Arunachal Pradesh gets red, Assam gets green, Manipur gets magenta, Meghalaya gets cyan, Mizoram gets brown, Nagaland gets pink, and Tripura gets purple. The first slice (Arunachal Pradesh) is exploded outward by 0.1 times the radius, and the fifth slice (Mizoram) is exploded by 0.2 times the radius — making it stand out even more. Each slice now also displays its exact percentage contribution to the total forest cover. …

A pie chart shows proportions, but the default look is often too plain. The textbook shows you how to customise it by adding a shadow, exploding a slice, and rotating the chart for better readability. You use the shadow parameter set to True to give the chart depth, and the explode parameter with a list of offset values to pull out a specific slice. The startangle parameter rotates the entire pie so the first slice starts at a given angle, and autopct formats the percentage labels displayed on each slice.

import matplotlib.pyplot as plt

subjects = ['Maths', 'Science', 'English', 'Hindi', 'SST']
marks = [85, 78, 92, 70, 88]
explode = [0, 0, 0.1, 0, 0] …