Seaborn Basics & Plot Levels

📦 Install + import convention

pip install seaborn

import seaborn as sns
import matplotlib.pyplot as plt

🧱 Two kinds of plotting functions

Level	Returns	Can combine with subplots?	Examples
Axes-level	a single Axes	Yes, pass ax=	scatterplot, lineplot, barplot, boxplot, histplot, heatmap
Figure-level	a whole Figure, manages its own subplots	No, has its own col=/row= faceting instead	relplot, catplot, displot, lmplot, pairplot, jointplot

# Axes-level: fits into an existing subplot grid
fig, axs = plt.subplots(1, 2)
sns.scatterplot(data=df, x="x", y="y", ax=axs[0])
sns.boxplot(data=df, x="cat", y="y", ax=axs[1])

# Figure-level: creates its own figure, use col=/row= for faceting instead of manual subplots
sns.relplot(data=df, x="x", y="y", col="category", kind="scatter")

Figure-level functions don't take ax=. If you need one plot inside an existing subplot grid, use the axes-level equivalent (relplot → scatterplot/lineplot, catplot → barplot/boxplot/etc, displot → histplot/kdeplot).

Figure-level = fastest way to facet by a category. Axes-level = needed for precise control or embedding in a custom subplot layout.

📊 Built-in sample datasets

sns.get_dataset_names()          # list available sample datasets
df = sns.load_dataset("tips")       # load one — great for testing/practice
df = sns.load_dataset("iris")
df = sns.load_dataset("titanic")

🧮 The core call shape

sns.scatterplot(data=df, x="col1", y="col2", hue="col3", size="col4", style="col5")

Param	Role
data=	the DataFrame
x=, y=	column names (strings), not raw arrays
hue=	color by category
size=	point/line size by value
style=	marker/linestyle by category

Always pass column names as strings + data=df, not df["col"] directly — this is what lets Seaborn auto-generate legends and labels correctly.

👁️ Display

plt.show()          # same as matplotlib, still required in scripts

🔗 Next

Seaborn Relational Plots · Seaborn Statistical Estimation & Data Handling