Online Seaborn Compiler

Seaborn is a Python library for statistical charts, built on Matplotlib. You pass it a pandas DataFrame and name the columns to plot, and seaborn handles grouping, colors, legends and summaries such as means and confidence intervals. Data scientists use it to explore data: to see how values are spread, compare groups and spot relationships between columns. This page is an online seaborn compiler: the code runs in your browser, so you can try it without installing anything. Run the example first, then paste any snippet below into a new cell to try it.

What the example does

sns.lineplot() draws a line through the points (1, 2), (2, 5) and (3, 12). With plain lists the axes have no labels. Pass a DataFrame as data and column names as x and y, and seaborn labels each axis with its column name. The call returns the Matplotlib Axes it drew on, which the cell displays because it is the last line. When several rows share an x value, lineplot() draws their mean and shades a 95% confidence interval around it.

Draw a scatter plot colored by group

hue colors each point by the value in another column and adds a legend. style and size do the same for the marker shape and size:

import numpy as np
import pandas as pd
import seaborn as sns

rng = np.random.default_rng(seed=0)
hours = rng.uniform(0, 10, 60)
df = pd.DataFrame({
    "hours": hours,
    "score": 50 + 4 * hours + rng.normal(0, 6, 60),
    "course": rng.choice(["Math", "History"], 60),
})
sns.scatterplot(data=df, x="hours", y="score", hue="course")

Show a distribution with a histogram

histplot() counts the values that fall in each bin. With hue, it draws one histogram per group, overlapping in the same axes, and kde=True adds a smoothed density curve for each. Pass multiple="stack" to stack them instead:

import numpy as np
import pandas as pd
import seaborn as sns

rng = np.random.default_rng(seed=1)
df = pd.DataFrame({
    "minutes": np.concatenate([rng.normal(30, 5, 200), rng.normal(45, 8, 200)]),
    "mode": ["Train"] * 200 + ["Bus"] * 200,
})
sns.histplot(data=df, x="minutes", hue="mode", kde=True, bins=20)

Compare groups with a box plot

Each box spans the middle half of a group's values, with a line at the median. The whiskers reach the furthest values within 1.5 box lengths of the box, and anything beyond them is drawn as a separate point, like Wednesday's 180 here:

import pandas as pd
import seaborn as sns

df = pd.DataFrame({
    "day": ["Mon"] * 5 + ["Tue"] * 5 + ["Wed"] * 5,
    "sales": [120, 135, 128, 150, 110, 140, 155, 160, 138, 149,
              90, 105, 98, 180, 101],
})
sns.boxplot(data=df, x="day", y="sales")

sns.barplot() takes the same arguments and draws each group's mean, with a 95% confidence interval as the error bar.

Draw a heatmap of correlations

df.corr() gives the correlation between each pair of columns, from -1 to 1. annot=True writes each value in its cell, and vmin=-1 and vmax=1 fix the color scale: with cmap="coolwarm", negative values are blue, positive ones red and values near 0 gray:

import numpy as np
import pandas as pd
import seaborn as sns

rng = np.random.default_rng(seed=2)
temp = rng.normal(20, 5, 100)
df = pd.DataFrame({
    "temperature": temp,
    "ice_cream": 3 * temp + rng.normal(0, 5, 100),
    "umbrellas": -2 * temp + rng.normal(0, 10, 100),
    "noise": rng.normal(0, 1, 100),
})
sns.heatmap(df.corr(), annot=True, fmt=".2f", cmap="coolwarm", vmin=-1, vmax=1)

Good to know

• relplot(), displot(), catplot(), pairplot() and jointplot() return a grid object, not an Axes, and on this page the cell prints the object instead of the chart. Add import matplotlib.pyplot as plt and end the cell with plt.show() to display it.
• Two calls such as scatterplot() and histplot() in one cell draw on the same Axes. For side-by-side plots, create fig, axes = plt.subplots(1, 2) and pass ax=axes[0] to one call and ax=axes[1] to the other.
• sns.load_dataset(), which the seaborn docs use for their examples, downloads the data from GitHub, so it needs an internet connection.