Matplotlib: Figure, Axes and Your First Chart

Matplotlib: Figure, Axes and Your First Chart

Matplotlib is Python’s original visualization library. Seaborn and pandas use it behind the scenes: understanding how it works will help you everywhere.

Figure and Axes

A Matplotlib chart is organized on two levels:

• the Figure (Figure): the “sheet” that contains everything;
• one or more Axes (Axes): plotting areas, each with its own x axis, y axis and title.

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(8, 4))   # an 8 x 4 inch figure, a single Axes

You then draw by calling the methods of ax. This is the object-oriented interface, which is the recommended one. There is a shorter interface (plt.plot(...), plt.title(...)) that acts on the “current axes”: handy for a quick test, but confusing as soon as there are several charts.

A first line

mois = ["janv", "févr", "mars", "avr", "mai", "juin"]
ca_2025 = [120, 135, 128, 150, 162, 170]
ca_2026 = [130, 142, 155, 149, 175, 190]

fig, ax = plt.subplots(figsize=(8, 4))
ax.plot(mois, ca_2025, marker="o", label="2025")            # one series
ax.plot(mois, ca_2026, marker="o", linestyle="--", label="2026")  # a second one

ax.set_title("Monthly revenue")               # title
ax.set_xlabel("Month")                        # x axis label
ax.set_ylabel("Revenue (k€)")                 # with the unit!
ax.legend()                                   # shows the labels
ax.grid(alpha=0.3)                            # subtle grid
plt.show()                                    # in a notebook, display is automatic

Each call to ax.plot adds a line (Line2D) to the Axes. The label parameter gives its name in the legend, which only appears if you call ax.legend().

Style

The most useful parameters of plot:

Parameter	Examples	Effect
color	"tab:blue", "#1f77b4"	color
marker	"o", "s", "^"	symbol on each point
linestyle	"-", "--", ":"	line type
linewidth	2	thickness

ax.set_ylim(0, None) forces the y axis to start at zero, which avoids visually exaggerating variations. ax.set(title=..., xlabel=..., ylabel=...) sets several properties at once.

A line connects the points: it suggests continuity. Reserve it for ordered data, typically time. To compare categories (regions, products), use bars (next lesson).

Saving

fig.savefig("ca_mensuel.png", dpi=150, bbox_inches="tight")
plt.close(fig)   # frees memory when you generate many figures

bbox_inches="tight" trims unnecessary margins; the format is inferred from the extension (.png, .svg, .pdf).

Checking a chart with code

Matplotlib objects can be queried, which makes it possible to test a chart:

ax.get_title()            # "Monthly revenue"
len(ax.get_lines())       # 2 lines
[t.get_text() for t in ax.get_legend().get_texts()]   # ['2025', '2026']

This is how the exercises in this course check your charts.

Key takeaways

• fig, ax = plt.subplots() then ax.* methods.
• Title, labels with units, legend: always.
• Line for time; savefig to export.