Seaborn Styling & Themes

🎨 set_theme — global one-liner

sns.set_theme()                                     # apply Seaborn's default look to everything after this
sns.set_theme(style="whitegrid", palette="pastel")      # common combo
sns.set_theme(style="darkgrid", context="talk")            # bigger fonts, for presentations

Call sns.set_theme() once at the top of a script/notebook — every plot after it (Seaborn AND plain matplotlib) inherits the styling.

🖼️ style — background/grid look

sns.set_style("darkgrid")      # default, gray background + white gridlines
sns.set_style("whitegrid")        # white background + gray gridlines
sns.set_style("dark")                # gray background, no grid
sns.set_style("white")                  # plain white, no grid — good for adding own annotations
sns.set_style("ticks")                     # white + tick marks, no grid

with sns.axes_style("whitegrid"):      # apply temporarily, only within this block
    sns.scatterplot(data=df, x="x", y="y")

🔍 context — scale for output medium

sns.set_context("paper")        # smallest — for print/publication
sns.set_context("notebook")        # default
sns.set_context("talk")               # larger — presentations
sns.set_context("poster")                # largest — posters

Use context to resize fonts/lines/markers for the actual output medium, without manually tweaking every font size.

🌈 Color palettes

sns.set_palette("pastel")
sns.set_palette("Set2")
sns.set_palette("viridis")
 
sns.color_palette("husl", 8)          # 8 evenly-spaced hues
sns.color_palette("coolwarm", as_cmap=True)   # get it as a matplotlib colormap object

sns.scatterplot(data=df, x="x", y="y", hue="cat", palette="Set2")   # per-plot override, no need to set globally

Palette type	Examples	Use for
Qualitative	"Set1", "Set2", "pastel", "husl"	Unordered categories
Sequential	"Blues", "viridis", "rocket"	Ordered/continuous data
Diverging	"coolwarm", "RdBu", "vlag"	Data with a meaningful midpoint

sns.color_palette("husl", 8)      # preview: run in Jupyter, auto-displays swatches
sns.palplot(sns.color_palette("husl", 8))   # explicit swatch display

🎨 Custom palette

custom = ["#e74c3c", "#3498db", "#2ecc71"]
sns.set_palette(custom)
sns.scatterplot(data=df, x="x", y="y", hue="cat", palette=custom)

🧵 despine — clean up borders

sns.despine()                          # removes top + right spines (default), matplotlib equivalent in Matplotlib Axes Configuration
sns.despine(left=True)                    # also remove left spine
sns.despine(offset=10, trim=True)            # offset spines outward + trim to data range

🔗 Next

Seaborn Statistical Estimation & Data Handling · Matplotlib Styling & Customization