Multivariate views#

In this notebook, we show a few examples of how to have plots with graphs of different types in a figure, like having a scatter plot with marginal distributions or even a multivariate plot with pair relationships of all properties in a table.

Because these plots involve managing subplots, they are all figure-level functions.

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
sns.set_theme()

Let’s load the same dataframe.

df = pd.read_csv("../../data/BBBC007_analysis.csv")
df.head()
area intensity_mean major_axis_length minor_axis_length aspect_ratio file_name
0 139 96.546763 17.504104 10.292770 1.700621 20P1_POS0010_D_1UL
1 360 86.613889 35.746808 14.983124 2.385805 20P1_POS0010_D_1UL
2 43 91.488372 12.967884 4.351573 2.980045 20P1_POS0010_D_1UL
3 140 73.742857 18.940508 10.314404 1.836316 20P1_POS0010_D_1UL
4 144 89.375000 13.639308 13.458532 1.013432 20P1_POS0010_D_1UL

Plotting joint and marginal distributions#

To have a joint distribution of two variableswith the marginal distributions on the sides, we can use jointplot.

sns.jointplot(data=df, x="aspect_ratio", y="area")

<seaborn.axisgrid.JointGrid at 0x20aa156d070>
../_images/04_Multivariate_views_7_1.png

As expected, it is possible to separate groups by passing a categorical property to the hue argument. This has an effect on the marginal distribution, turning them from histogram to kde plots.

sns.jointplot(data=df, x="aspect_ratio", y="area", hue = 'file_name')

<seaborn.axisgrid.JointGrid at 0x20aa1574ca0>
../_images/04_Multivariate_views_9_1.png

Plotting many distributions at once#

The above examples displayed a plot with relationship between two properties. This can be further expanded with the pairplot function

sns.pairplot(data=df)

<seaborn.axisgrid.PairGrid at 0x20aa76121f0>
../_images/04_Multivariate_views_12_1.png 
sns.pairplot(data=df, hue="file_name")

<seaborn.axisgrid.PairGrid at 0x20aa892e070>
../_images/04_Multivariate_views_13_1.png

If you have too many points, displaying every single point may yield graphs too poluted. An alternative visualization in this case could be a 2D histogram plot. We can do that by changing the kind argument to “hist”.

sns.pairplot(data=df, hue="file_name", kind = "hist")

<seaborn.axisgrid.PairGrid at 0x20abbf7d4f0>
../_images/04_Multivariate_views_15_1.png

Exercise#

You may have noticed that the pairplot is redundant in some plots because the upper diagonal displays the same relationships rotated.

Redraw the pairplot to display only the lower diagonal of the plots.

Hint: explore the properties of the pairplot