Charts with Matplotlib

Why this matters

A good chart shows a pattern in seconds that a table of numbers hides. Matplotlib is Python’s most widely used plotting library, and many other tools (pandas plotting, seaborn) are built on it. This lesson runs on your own computer, because charts open in a window or are saved as picture files.

Install the extras once with pip install -r requirements-tracks.txt.

The idea

import matplotlib.pyplot as plt

fig, ax = plt.subplots()           # a Figure (the canvas) and an Axes (one chart on it)
ax.plot([1, 2, 3, 4], [10, 12, 9, 15])      # x values, then y values: a line chart
ax.set_title("Daily steps")
ax.set_xlabel("Day")
ax.set_ylabel("Steps (hundreds)")
plt.show()                          # opens a window (on your own computer)

• fig is the whole picture; ax is the chart inside it. You draw on the Axes and label it with set_title, set_xlabel and set_ylabel.
• A bar chart is ax.bar(labels, heights). A scatter plot is ax.scatter(xs, ys). A histogram is ax.hist(values).
• fig.savefig("steps.png") writes the chart to an image file instead of (or as well as) showing it. Do this in scripts that run without a screen.

Always label what the chart shows: a title, axis labels with units, and a legend (ax.legend(), with label= on each plot) when there are several lines.

pandas can draw straight from a table too (df.plot(...)), and it returns the same Axes object.

Try it

Plot the numbers 1 to 5 against their squares and show the chart with plt.show(). Then change it to a bar chart.

Your turn

The checks run without opening a window and inspect the chart you built. Write:

• line_chart(days, values): a line chart of values against days, titled "Daily steps", with the x label "Day" and the y label "Steps". Return the Axes.
• bar_chart(labels, counts): a bar chart with one bar per label, titled "Votes". Return the Axes.
• save_chart(ax, path): save the chart to the file path as an image.

Recap

• fig, ax = plt.subplots() makes the canvas and one chart; you draw and label on ax.
• plot (lines), bar, scatter and hist are the everyday chart types.
• plt.show() shows a window; fig.savefig(path) writes a file.
• Titles, axis labels and units are part of the chart, not extras.

Common mistakes

• Forgetting plt.show() or savefig in a script, so nothing appears.
• Mixing the older plt.plot(...) style with ax.plot(...) and drawing on the wrong chart. Pick one; the Axes style scales better.
• Unlabelled axes. A chart without units is a puzzle, not a message.