One of the most important parts of data analysis is in creating a great visualization to immediately convey the underlying meaning in the data. Data visualization is effective, as we humans are visual creatures and have evolved to be able to discern meaning when information is laid out in a way that our brain can interpret almost immediately when the impulses from the retina hit the brain.

Over the years, there has been significant research that has resulted in many effective visualization techniques to convey specific patterns in data. These patterns have been implemented in visualization libraries, and pandas is designed to utilize these and make their use very simple.

This chapter will cover several of these techniques, primarily focusing on matplotlib, and many of the common visualizations. We will do this in three steps. The first introduces the general concepts of programming visualizations with pandas, emphasizing the process of creating time-series charts. During this, we will dive into the techniques of labeling axes and creating legends, colors, line styles, and markers.

The second step will focus on the many types of data visualizations commonly used in pandas and data analysis, including:

• Showing relative differences with bar plots
• Picturing distributions of data with histograms
• Depicting distributions of categorical data with box and whisker charts
• Demonstrating cumulative totals with area plots
• Relationships between two variables with scatter plots
• Estimates of distribution with the kernel density plot
• Correlations between multiple variables with the scatter plot matrix
• Strengths of relationships in multiple variables with heatmaps

The final step will examine creating composite plots by dividing plots into subparts to be able to render multiple plots within a single graphical canvas. This will help the viewer of the visualization to relate different sets of data at a glance.