Matplotlib Visualization Guide

Problem Addressed

In matplotlib charting, issues are often not about drawing a plot, but making plots readable and code maintainable: stateful pyplot can mix axes in multi-plot workflows; labels, legends, and ticks can overlap; colormaps can be poorly matched to data; exports to PNG, PDF, and SVG can have unstable margins and resolution. This skill turns those concerns into concrete choices.

How the Skill Works

It is organized around the Figure, Axes, Artist, and Axis hierarchy, and distinguishes the quick pyplot interface from the object-oriented API. Production code should prefer fig, ax = plt.subplots() for explicit axis control. It covers line, scatter, bar, histogram, heatmap, contour, box, and violin plots, along with rcParams, stylesheets, colormaps, annotations, subplots layouts, dpi, bbox_inches='tight', and transparent backgrounds. For 3D plots, animation, and large datasets, it also flags tradeoffs such as rasterized=True, downsampling, and blitting.

Boundaries and Caveats

This skill fits Python, Jupyter, Tkinter, Qt, or wxPython workflows for static, animated, or interactive charts, and pairs well with NumPy, Pandas, and Seaborn. Keep in mind that figsize is in inches, not pixels; close figures with plt.close(fig) to avoid memory growth; handle font warnings via font.sans-serif; and prefer semantically appropriate colormaps over rainbow scales for classification or diverging data.