Libraries
First, we need to load a few libraries:
- seaborn: for creating the scatterplot
- matplotlib: for displaying the plot
- pandas: for data manipulation
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as snsDataset
Since scatter plot are made for visualizing relationships between two numerical variables, we need a dataset that contains at least two numerical columns.
Here, we will use the iris dataset that we load directly from the gallery:
path = 'https://raw.githubusercontent.com/holtzy/The-Python-Graph-Gallery/master/static/data/iris.csv'
df = pd.read_csv(path)
df.head()| sepal_length | sepal_width | petal_length | petal_width | species | |
|---|---|---|---|---|---|
| 0 | 5.1 | 3.5 | 1.4 | 0.2 | setosa |
| 1 | 4.9 | 3.0 | 1.4 | 0.2 | setosa |
| 2 | 4.7 | 3.2 | 1.3 | 0.2 | setosa |
| 3 | 4.6 | 3.1 | 1.5 | 0.2 | setosa |
| 4 | 5.0 | 3.6 | 1.4 | 0.2 | setosa |
Color and width
You can custom the appearance of the regression fit in a scatterplot built with seaborn thanks to the line_kws argument.
This argument accepts a dictionary of argument names and their values. These values will be exclusively used to style the line. For example, line_kws={"color" : "red", "linewidth" : 1.5} will render the line red and with a width of 1.5.
fig, ax = plt.subplots(figsize=(8, 6))
sns.regplot(
x=df["sepal_length"],
y=df["sepal_width"],
line_kws={"color": "red", "linewidth": 1.5},
ax=ax
)
plt.show()
Opacity
You can also custom the opacity of the line with the alpha value:
fig, ax = plt.subplots(figsize=(8, 6))
sns.regplot(
x=df["sepal_length"],
y=df["sepal_width"],
line_kws={
"color": "r",
"alpha": 0.4
},
ax=ax
)
plt.show()
Line width and style
You can also custom the line width and style with the linewidth and linestyle values:
fig, ax = plt.subplots(figsize=(8, 6))
sns.regplot(
x=df["sepal_length"],
y=df["sepal_width"],
line_kws={
"color": "darkred",
"alpha": 0.4,
"lw": 5,
"ls": "--"
},
ax=ax
)
plt.show()
Going further
This post explains how to customize the appearance of a regression fit in a scatter plot with seaborn.
You might be interested in more advanced examples on:
