Place a Matplotlib Legend Outside the Plot Area

You will learn how Matplotlib positions legends, how to anchor a legend outside an axes, and how to make a large legend more compact with options such as fontsize, ncol, and spacing settings. You will also learn how to save an external legend without clipping it or shrinking the plotted area.

A Matplotlib legend is an artist that describes plotted objects with labels, usually lines, markers, or bars. By default, calling ax.legend() places it inside the axes.

To move a legend outside the plotted rectangle, use two arguments together:

• loc: chooses the point on the legend box to use as an anchor.
• bbox_to_anchor: supplies the destination coordinates for that anchor point.

For example, loc="upper left" and bbox_to_anchor=(1.02, 1) mean: place the legend's upper-left corner just to the right of the axes' upper-right area.

A legend outside the axes still needs space on the figure canvas. You can choose between two common approaches:

1. Keep the axes position unchanged and save with bbox_inches="tight". The exported image expands its bounding box to include the external legend.
2. Reserve space in the figure layout. This prevents overlap or clipping in an interactive window, but it usually reduces the available axes area unless you make the overall figure wider.

For many legend entries, decreasing fontsize is useful, but it can make labels hard to read. A multi-column legend and smaller spacing often improve compactness more effectively.

Think of the axes as a framed photograph and the legend as a caption card.

• By default, Matplotlib puts the caption card on top of the photograph.
• loc chooses which corner of the card you hold.
• bbox_to_anchor tells you where to place that held corner.

If you place the card beyond the photograph's frame, the overall page must either become wider or be cropped carefully when it is exported. Reducing fontsize makes the card smaller, but it does not create unlimited room.

Use bbox_to_anchor with loc to place a legend on the right side of an axes:

import matplotlib.pyplot as plt

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

ax.plot([1, 2, 3], [2, 4, 3], label="Sales")
ax.plot([1, 2, 3], [1, 3, 5], label="Profit")

ax.legend(
    loc="upper left",
    bbox_to_anchor=(1.02, 1),
    fontsize=8
)

ax.set_xlabel("Month")
ax.set_ylabel("Amount")
plt.show()

bbox_to_anchor=(1.02, 1) uses axes-relative coordinates by default:

• 0, 0 is the lower-left of the axes.
• 1, 1 is the upper-right of the axes.
• 1.02, 1 is slightly to the right of the upper-right corner.

For a legend with many labels, use columns and compact spacing:

Consider this code:

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(7, 4))

line_a, = ax.plot([0, 1, 2], [1, 2, 1], label="Experiment A")
line_b, = ax.plot([0, 1, 2], [2, 1, 2], label="Experiment B")

legend = ax.legend(
    loc="center left",
    bbox_to_anchor=(1.02, 0.5),
    fontsize=8
)

fig.savefig("results.png", bbox_inches="tight", dpi=150)

Execution trace:

1. plt.subplots() creates a figure and one axes.
2. Each ax.plot() call creates a line. The label values are stored for the legend.
3. ax.legend() builds legend entries from the labeled lines.
4. loc="center left" selects the center of the legend's left edge as its anchor point.
5. bbox_to_anchor=(1.02, 0.5) places that point just right of the axes and halfway up its height.

External legends are useful when the plotted data should remain unobstructed:

• Monitoring dashboards: several time-series lines need a readable plot area and a list of service names beside it.
• Scientific charts: experiment names can be long, so an internal legend would cover data points.
• Financial reports: multiple funds or market indicators need labels without hiding price trends.
• Data exports: a script generates PNG or PDF reports where the legend must be included reliably.
• Dense comparison charts: 10–20 categories may need a multi-column legend below or beside the plot.

In production plotting code, developers commonly keep the legend configuration in one place so all generated charts look consistent.

LEGEND_OPTIONS = {
    "loc": "upper left",
    "bbox_to_anchor": (1.02, 1),
    "fontsize": 8,
    "frameon": False,
    "labelspacing": 0.3,
}

ax.legend(**LEGEND_OPTIONS)

Useful real-project patterns include:

• Only label needed artists: give label only to lines that should appear in the legend. This prevents duplicate or noisy entries.
• Use figure-level legends for multiple axes: fig.legend(...) can create one shared legend for a multi-panel figure.
• Export safely: use fig.savefig(path, bbox_inches="tight") for reports that contain annotations or legends outside axes.
• Make output dimensions intentional: if an external legend must be visible in an application window, increase figsize rather than allowing the axes to become too small.
• Use readable labels: shortening verbose names can be better than making the font extremely small.

Expecting an external legend to need no space

An external legend is outside the axes, but it is still part of the figure. If the figure is not wide enough, it may be clipped or overlap another element.

# May be clipped in a saved image without bbox_inches="tight"
ax.legend(loc="upper left", bbox_to_anchor=(1.02, 1))
fig.savefig("plot.png")

Prefer:

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

Using bbox_to_anchor without understanding loc

This does not mean “put the entire legend at this coordinate.” It means “put the point selected by loc at this coordinate.”

# The legend's lower-right corner is anchored at the axes upper-right corner.
ax.legend(loc="lower right", bbox_to_anchor=(1, 1))

Choose loc="upper left" for a legend placed to the right of an axes.

Making the font too small

ax.legend(fontsize=3)

# Legend outside the right edge of one axes
ax.legend(
    loc="upper left",
    bbox_to_anchor=(1.02, 1),
    fontsize=8
)

# Centered vertically on the right
ax.legend(
    loc="center left",
    bbox_to_anchor=(1.02, 0.5)
)

# Compact legend for many entries
ax.legend(
    loc="upper left",
    bbox_to_anchor=(1.02, 1),
    fontsize=7,
    ncol=2,
    labelspacing=0.3,
    handlelength=1.2,
    frameon=False
)

# Save an external legend without cropping it
fig.savefig("chart.png", bbox_inches="tight", dpi=150)

Rules to remember:

• loc chooses the legend point to anchor.
• bbox_to_anchor chooses where that point goes.
• Coordinates are relative to the axes by default.
• An external legend needs canvas space when displayed or exported.
• Use columns and spacing options before making text too small.

How do I move a Matplotlib legend outside the plot on the right?

Use loc="upper left" with bbox_to_anchor=(1.02, 1) in ax.legend().

Does bbox_to_anchor change the size of the axes?

No. It changes the legend position. However, the figure may need additional canvas space to display or save the legend without clipping.

How do I reduce the legend font size in Matplotlib?

Pass a number to fontsize, such as ax.legend(fontsize=8).

Why is my legend missing from the saved PNG?

It is probably outside the normal figure boundary. Save with fig.savefig("plot.png", bbox_inches="tight").

How can I fit 20 legend labels into less space?

Try ncol=2 or more, shorten labels, reduce labelspacing, and use a modestly smaller fontsize.

Should I use ax.legend() or fig.legend()?

Use ax.legend() for one axes. Use fig.legend() when one legend should describe data across multiple axes.

• Matplotlib axes and figure objects — Helps distinguish an individual plotting area from the full drawing canvas.
• Matplotlib labels and annotations — Legends use artist labels and share similar placement concerns with text annotations.
• Matplotlib savefig() — Important for exporting figures that contain content outside the axes.
• Matplotlib layout management — tight_layout, constrained layout, and manual margins affect how figure elements fit together.
• Multi-panel Matplotlib figures — Figure-level legends are especially useful when several axes share labels.
• Data visualization readability — Label length, font size, color, and chart dimensions affect whether a chart remains understandable.