Matplotlib Cheat Sheet

01Setup & Createimport · figure · axes

import matplotlib.pyplot as plt★
The one import you always write.
import numpy as np
Data almost always comes in as arrays.
fig, ax = plt.subplots()★start here
One Figure + one Axes. The modern default entry point.
fig, ax = plt.subplots(figsize=(8,5))★
Size in inches (width, height).
fig, axs = plt.subplots(2, 3)★
A 2×3 grid — axs is a NumPy array.
fig, axs = plt.subplots(2, 2, layout='constrained')
Auto-spaces subplots so nothing overlaps.
plt.plot(x, y)
Quick pyplot form — auto-creates fig + axes for you.

02Basic / Pairwiseplots of (x, y)

03Statisticaldistributions

04Gridded & Fields2-D arrays Z, images

05Labels & Titlewords on the plot

ax.set_title("...")★
Title above this Axes.
ax.set_xlabel("...")★
x-axis label; set_ylabel for y.
ax.set(xlabel="t", ylabel="v", title="...")★
Set many properties in one call.
ax.legend()★
Needs label= on each plot call first.
fig.suptitle("...")
One title over the whole Figure.
ax.annotate("peak", xy=p, xytext=t, arrowprops={...})
Text with an arrow pointing at a datapoint.
ax.text(x, y, "note")
Free text at data coordinates.

06Limits, Ticks & Scalethe axis itself

ax.set_xlim(0, 10)★
View range; set_ylim for y.
ax.set_xticks([0,5,10])★
Where ticks sit; add a 2nd list for labels.
ax.tick_params(axis='x', rotation=45)
Rotate / resize tick labels.
ax.set_xscale('log')★
Also 'symlog', 'logit'.
ax.grid(True)★
Toggle the grid; which='both' incl. minor.
ax.axhline(0) · ax.axvline(x)
Full-width reference lines.
ax.invert_yaxis()
Flip an axis direction.
ax.xaxis.set_major_locator(MultipleLocator(5))
Fine tick control via Locators / Formatters.

07Colors, Lines & Markersstyling one artist

ax.plot(..., color='C0')★
'C0'…'C9' = the default color cycle.
linewidth=2 · linestyle='--'★
Aliases lw / ls. Styles: - -- -. :
marker='o' · markersize=8★
Markers: o s ^ v . * + x D
alpha=0.5
Transparency, 0 (clear) → 1 (solid).
label="series A"★
Feeds ax.legend().
zorder=3
Higher = drawn on top.
# color forms: 'red' 'r' '#1f77b4' (.2,.4,.6) '0.5'
Name, 1-letter, hex, RGB tuple, or gray level.

08Colormaps & Colorbardata → color

ax.imshow(Z, cmap='viridis')★
Perceptually uniform, colorblind-safe default.
# sequential viridis plasma magma cividis
Low→high ordered data.
# diverging coolwarm RdBu bwr
Data with a meaningful center (0).
# qualitative tab10 Set2 Pastel1
Unordered categories.
fig.colorbar(im, ax=ax)★
The key relating color back to value.
vmin=0, vmax=1
Clamp the value→color range.
norm=LogNorm()
Non-linear mapping (e.g. log). Append _r to reverse any cmap.
matplotlib.colormaps['viridis']
Get a colormap object by name. cm.get_cmap & register_cmap were removed — use matplotlib.colormaps[...] and matplotlib.colormaps.register(...).

09Subplots & Layoutmany Axes, one Figure

fig, axs = plt.subplots(2, 3)★
Regular grid → index axs[row, col].
plt.subplots(..., layout='constrained')★
Modern auto-spacing. Prefer over tight_layout().
fig.subplot_mosaic("AB;CC")★
Named, uneven layouts by ASCII art → dict of axes.
plt.subplots(..., sharex=True)
Share axis range/ticks across panels.
ax2 = ax.twinx()
Second y-axis sharing the same x.
iax = ax.inset_axes([.6,.6,.35,.35])
A small plot inside the plot.
fig.add_gridspec(3, 3)
Full manual control of a grid.

10Save & Showget it out

plt.show()★
Render window / inline. Call once, at the end.
fig.savefig("f.png", dpi=300)★
Raster export; bump dpi for print quality.
fig.savefig("f.pdf", bbox_inches='tight')★
Vector (pdf/svg); tight trims whitespace.
fig.savefig("f.svg", transparent=True)
No background — good for slides.
plt.close(fig)
Free the figure's memory (batch jobs!).

11Style & Defaultsthe global look

123-D, Polar & Motionbeyond flat

ax = fig.add_subplot(projection='3d')★
Opt into a 3-D Axes.
ax.plot_surface(X, Y, Z)
A 3-D surface; also scatter(x,y,z).
fig.add_subplot(projection='polar')
Radial / angular plots.
ax.set_aspect('equal')
Equal x/y scale — circles stay round.
from matplotlib.animation import FuncAnimation
Frame-by-frame animation → .save("a.gif").

13With pandasthe common shortcut

★The Two Interfacesread this once

# explicit / OO (recommended)prefer
You hold fig/ax and call methods on them.
fig, ax = plt.subplots(); ax.plot(x,y); ax.set_title(...)
Clear, scales to complex/multi-panel figures.
# implicit / pyplot (state-machine)
pyplot tracks a "current" Axes for you.
plt.plot(x,y); plt.title(...)
Fewer keystrokes; fine for quick throwaway plots.
plt.plot() ≡ plt.gca().plot()
pyplot is a thin wrapper: plt.title ↔ ax.set_title.

15Gotchaswhere hours go

Axes ≠ Axisconfusing
Axes = the whole plot; Axis = one x or y line.
ax.legend() shows nothing
You forgot label= on the plot calls.
from pylab import *deprecated
Strongly discouraged — pollutes the namespace.
axs[0] vs axs[0][1]
subplots(2,3) gives a 2-D array — index both.
too many open figures
Loops leak memory — plt.close(fig) each time.
plt.show() then savefig = blank
show() can clear the figure; save first.
imshow origin is top-left
Row 0 is at the top; use origin='lower' to flip.

★The fmt String & Colors'[marker][line][color]'

'ro--'
red o-markers, -- dashed line.
'k.:'
k black, . point markers, : dotted.
# line - solid -- dash -. dashdot : dot
The 4 line styles.
# marker o s ^ v < > . * + x D p
Common marker glyphs.
# color b g r c m y k w
1-letter colors (blue…white).
'C0'..'C9'
The property-cycle colors, in order.
'#1f77b4' (0.1,0.2,0.5) 'tab:blue' '0.5'
Hex · RGB tuple · tableau name · gray level.

matplotlib cheat sheet