Plotting with matplotlib — Python Module 10 | LeadAfrik Public Economics Hub

matplotlib is the foundation plotting library for Python. Its API is two-track: a pyplot 'state machine' interface that mimics MATLAB, and an explicit object-oriented interface that scales to complex figures. Modern practice uses the explicit fig/ax pattern even for simple plots.

The fig/ax pattern

python

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(8, 5))
ax.plot(bankrates['month'], bankrates['lending_rate'], label='Lending')
ax.plot(bankrates['month'], bankrates['deposit_rate'], label='Deposit')
ax.set_xlabel('Month')
ax.set_ylabel('Rate (%)')
ax.set_title('Kenyan commercial bank rates, 2023-2024')
ax.legend()
plt.show()

Common plot types

python

ax.plot(x, y)              # line
ax.scatter(x, y)           # scatter
ax.bar(x, heights)         # bar
ax.hist(values, bins=20)   # histogram
ax.boxplot(arrays)         # boxplot

Multiple subplots

python

fig, axes = plt.subplots(1, 2, figsize=(12, 4))
axes[0].plot(x, y1)
axes[0].set_title('Lending')
axes[1].plot(x, y2)
axes[1].set_title('Deposit')
fig.tight_layout()

Pandas .plot()

Pandas wraps matplotlib in a more concise API. For exploratory plots, df.plot() is hard to beat.

python

bankrates.plot(x='month', y=['lending_rate', 'deposit_rate'])

Styling — keep it minimal

Resist the urge to over-style. The defaults in modern matplotlib are good. Add labels, a title, and a legend; resize for the medium; export to PNG or SVG.

When to reach for seaborn or plotly

Seaborn is matplotlib with smarter statistical defaults — quicker for distributions, regressions, and faceted plots. Plotly is interactive and ideal for dashboards. Both build on top of matplotlib for the underlying engine.

Exercise

Plot a line chart of bankrates['month'] vs bankrates['lending_rate'].