Visualization of Data with Pie Charts in Matplotlib | by Diana Rozenshteyn | Oct, 2024

Whereas engaged on my Grasp’s Thesis titled “Elements Related to Impactful Scientific Publications in NIH-Funded Coronary heart Illness Analysis”, I’ve used various kinds of pie charts for example among the key findings from the database evaluation.

A pie chart might be an efficient selection for knowledge visualization when a dataset comprises a restricted variety of classes representing elements of a complete, making it well-suited for displaying categorical knowledge with an emphasis on evaluating the relative proportions of every class.

On this article, I’ll exhibit tips on how to create 4 various kinds of pie charts utilizing the identical dataset to offer a extra complete visible illustration and deeper perception into the info. To attain this, I’ll use Matplotlib, Python’s plotting library, to show pie chart visualizations of the statistical knowledge saved within the dataframe. In case you are not acquainted with Matplotlib library, a very good begin is Python Knowledge Science Handbook by Jake VanderPlas, particularly chapter on Visualization with Matplotlib and matplotlib.org.

First, let’s import all the required libraries and extensions:

Subsequent, we’ll put together the CSV file for processing:

The mini dataset used on this article highlights the highest 10 journals for coronary heart illness analysis publications from 2002 to 2020 and is a component of a bigger database collected for the Grasp’s Thesis analysis. The columns “Feminine,” “Male,” and “Unknown” signify the gender of the primary creator of the revealed articles, whereas the “Complete” column displays the overall variety of coronary heart illness analysis articles revealed in every journal.

For smaller datasets with fewer classes, a pie chart with exploding slices can successfully spotlight a key class by pulling it out barely from the remainder of the chart. This visible impact attracts consideration to particular classes, making them stand out from the entire. Every slice represents a portion of the overall, with its measurement proportional to the info it represents. Labels might be added to every slice to point the class, together with percentages to point out their proportion to the overall. This visible method makes the exploded slice stand out with out shedding the context of the complete knowledge illustration.

The identical exploding slices method might be utilized to all different entries within the pattern dataset, and the ensuing charts might be displayed inside a single determine. This sort of visualization helps to focus on the over illustration or underneath illustration of a selected class throughout the dataset. Within the instance offered, presenting all 10 charts in a single determine reveals that not one of the prime 10 journals in coronary heart illness analysis revealed extra articles authored by ladies than males, thereby emphasizing the gender disparity.

A variation of the pie chart, generally known as a donut chart, may also be used to visualise knowledge. Donut charts, like pie charts, show the proportions of classes that make up a complete, however the middle of the donut chart may also be utilized to current extra knowledge. This format is much less cluttered visually and might make it simpler to match the relative sizes of slices in comparison with a typical pie chart. Within the instance used on this article, the donut chart highlights that among the many prime 10 journals for coronary heart illness analysis publications, the American Journal of Physiology, Coronary heart and Circulatory Physiology revealed probably the most articles, accounting for 21.8%.

We are able to improve the visualization of extra data from the pattern dataset by constructing on the earlier donut chart and making a nested model. The add_artist() technique from Matplotlib’s determine module is used to include any extra Artist (similar to figures or objects) into the bottom determine. Just like the sooner donut chart, this variation shows the distribution of publications throughout the highest 10 journals for coronary heart illness analysis. Nevertheless, it additionally contains a further layer that exhibits the gender distribution of first authors for every journal. This visualization highlights {that a} bigger proportion of the primary authors are male.

In conclusion, pie charts are efficient for visualizing knowledge with a restricted variety of classes, as they permit viewers to rapidly perceive an important classes or dominant proportions at a look. On this particular instance, using 4 various kinds of pie charts gives a transparent visualization of the gender distribution amongst first authors within the prime 10 journals for coronary heart illness analysis publications, based mostly on the 2002 to 2020 mini dataset used on this examine. It’s evident {that a} greater proportion of the publication’s first authors are males, and not one of the prime 10 journals for coronary heart illness analysis revealed extra articles authored by females than by males throughout the examined interval.

Jupyter Pocket book and dataset used for this text might be discovered at GitHub

Thanks for studying,

Diana

Observe: I used GitHub embeds to publish this text.