Skip to content

Matplotlib in Practice

Matplotlib is Python’s foundational plotting library. Most other plotting libraries (seaborn, pandas plotting, etc.) are built on top of it.

What Matplotlib is (and what it isn’t)

Matplotlib:

  • Produces static plots (PNG, PDF, SVG)
  • Gives you full control over plots
  • Works directly with NumPy arrays and pandas Series/DataFrames

Matplotlib is not:

  • Automatically “pretty”
  • Declarative (you build plots step by step)
  • Interactive by default (that’s a separate layer)

Think of Matplotlib as a drawing API for data.

The core mental model (very important)

Matplotlib has a hierarchy:

Figure
 └── Axes
      ├── Lines
      ├── Bars
      ├── Text
      └── Ticks
  • Figure → the whole image (canvas)
  • Axes → a single plot (x/y area)
  • You draw on Axes

Most confusion comes from mixing these up.

Importing Matplotlib (standard way)

import matplotlib.pyplot as plt

This plt interface is a state machine that keeps track of the “current” figure and axes.

Your first plot (line plot)

import matplotlib.pyplot as plt
import numpy as np

x = np.arange(0, 10)
y = x ** 2

plt.plot(x, y)
plt.show()

What happened:

  • plt.plot() created a figure + axes (implicitly)
  • plt.show() displayed it

This is the quick-and-dirty approach.

fig, ax = plt.subplots()

ax.plot(x, y)
plt.show()

Why this is better:

  • You explicitly control where things are drawn
  • Required for multiple plots, customization, reuse

Rule of thumb:

Use fig, ax = plt.subplots() for anything non-trivial.

Labeling your plot (always do this)

fig, ax = plt.subplots()

ax.plot(x, y)
ax.set_title("y = x²")
ax.set_xlabel("x")
ax.set_ylabel("y")

plt.show()

A plot without labels is unfinished.

Styling lines

ax.plot(x, y,
        color="red",
        linestyle="--",
        linewidth=2,
        marker="o")

Common options:

  • color
  • linestyle (-, --, :)
  • linewidth
  • marker

Multiple lines on the same axes

fig, ax = plt.subplots()

ax.plot(x, x, label="y = x")
ax.plot(x, x**2, label="y = x²")
ax.plot(x, x**3, label="y = x³")

ax.legend()
plt.show()

Legend works by matching label=.

Scatter plots (x vs y points)

rng = np.random.default_rng(0)
x = rng.normal(size=100)
y = rng.normal(size=100)

fig, ax = plt.subplots()
ax.scatter(x, y)
plt.show()

With color and size:

ax.scatter(x, y, c=y, s=50, alpha=0.7)

Bar plots (categorical data)

categories = ["A", "B", "C"]
values = [10, 15, 7]

fig, ax = plt.subplots()
ax.bar(categories, values)
plt.show()

Horizontal bars:

ax.barh(categories, values)

Histograms (distributions)

data = rng.normal(size=1000)

fig, ax = plt.subplots()
ax.hist(data, bins=30)
plt.show()

Common options:

  • bins
  • density=True
  • alpha

Plotting pandas data (very common)

Matplotlib works directly with pandas.

import pandas as pd

df = pd.DataFrame({
    "day": ["Mon", "Tue", "Wed", "Thu", "Fri"],
    "sales": [10, 12, 9, 14, 15]
})

fig, ax = plt.subplots()
ax.plot(df["day"], df["sales"])
plt.show()

Or via pandas’ wrapper:

df.plot(x="day", y="sales")
plt.show()

(Still Matplotlib underneath.)

Multiple subplots (one figure, many axes)

fig, axes = plt.subplots(2, 2)

axes[0, 0].plot(x, x)
axes[0, 1].plot(x, x**2)
axes[1, 0].plot(x, x**3)
axes[1, 1].plot(x, np.sqrt(x))

plt.tight_layout()
plt.show()

Key points:

  • axes is an array
  • Index it like NumPy

Controlling figure size and resolution

fig, ax = plt.subplots(figsize=(8, 4), dpi=100)

Important for:

  • Publications
  • PDFs
  • Presentations

Axis limits and scales

ax.set_xlim(0, 10)
ax.set_ylim(0, 100)

Log scale:

ax.set_yscale("log")

Ticks and formatting

ax.set_xticks([0, 5, 10])
ax.set_xticklabels(["start", "middle", "end"])

Rotate labels:

plt.setp(ax.get_xticklabels(), rotation=45)

Annotations and text

ax.text(5, 50, "Important point")

Arrow annotation:

ax.annotate("Peak",
            xy=(7, 49),
            xytext=(4, 70),
            arrowprops=dict(arrowstyle="->"))

Saving figures (very important)

fig.savefig("plot.png", dpi=300, bbox_inches="tight")
fig.savefig("plot.pdf")

Always save before plt.show() in scripts.

A realistic end-to-end example

import pandas as pd
import matplotlib.pyplot as plt

sales = pd.DataFrame({
    "date": pd.date_range("2026-01-01", periods=7),
    "units": [10, 12, 9, 14, 15, 11, 13]
})

fig, ax = plt.subplots(figsize=(8, 4))

ax.plot(sales["date"], sales["units"], marker="o")
ax.set_title("Daily Sales")
ax.set_xlabel("Date")
ax.set_ylabel("Units Sold")

plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

This pattern covers most real-world plots.

Common Matplotlib mistakes

  • Mixing plt.plot() and ax.plot() randomly
  • Forgetting labels and legends
  • Overusing defaults
  • Fighting Matplotlib instead of learning the hierarchy
  • Using loops when vectorized plotting works

The most important mindset shift

Matplotlib is imperative:

“Draw this, then draw that, then adjust this.”