Open
Description
Feature Type
-
Adding new functionality to pandas
-
Changing existing functionality in pandas
-
Removing existing functionality in pandas
Problem Description
I wish I could use the scatter_matrix function to make hexbin plots
Feature Description
Add a new parameter nondiagonal
to scatter_matrix
to support hexbin
instead of scatter
def scatter_matrix(...):
...
if nondiagonal=="scatter":
ax.scatter(...)
if nondiagonal=="hexbin":
ax.hexbin(...)
...
Alternative Solutions
Seaborn pairplot using kind = hist
. Link: https://seaborn.pydata.org/generated/seaborn.pairplot.html
Additional Context
The general consensus on pandas.plotting.scatter_matrix
is that "seaborn does this better" (ref 1, ref 2). Nevertheless, the functionality was kept, and the changes to support this feature are a few lines of code.
The quality of the plots can improve significantly. Here's an example with scatter:
# Copy of example from scatter_matrix docs page, but increase data frame rows from 1k to 10k:
import pandas as pd
import numpy as np
df = pd.DataFrame(np.random.randn(1_0_000, 4), columns=['A','B','C','D'])
Same example with hexbin: