Display multiple graphics with Seaborn

Question

I have a csv file that I open with Pandas. It looks like this:

"Patient","Visit","Feature1","Feature2","Feature3"
"P1",0,?,?,?
"P1",1,?,?,?
"P2",0,?,?,?
"P2",1,?,?,?
"P2",2,?,?,?
"P3",0,?,?,?
"P3",1,?,?,?

? being a numerical value.

I know how to display one feature at a time (because they have various ranges) and separated per patient:

seaborn.stripplot(y="Feature1", x="Patient", data=features)

I would like to display ALL the plots inside a single image, like a mapping. How can I do it?

Answer 1

Seaborn usually prefers the "long form" of the dataframe, which can be obtained via pandas' melt.

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

df = pd.DataFrame({"Patient": ["P1", "P1", "P2", "P2", "P2", "P3", "P3"],
                   "Visit": [0, 1, 0, 1, 2, 0, 1],
                   "Feature1": np.random.normal(10, 2, 7),
                   "Feature2": np.random.normal(12, 2, 7),
                   "Feature3": np.random.normal(15, 2, 7)})
df_melted = df.melt(value_vars=["Feature1", "Feature2", "Feature3"], id_vars=["Patient", "Visit"],
                    var_name="Feature", value_name="Value")
sns.stripplot(data=df_melted, x="Patient", y="Value", hue="Feature", dodge=True)

plt.tight_layout()
plt.show()

To get each feature in separate subplots, you can use catplot on the melted dataframe (use sharey=True if all features are in about the same range; use col_wrap=3 to start a new row every 3 columns):

sns.catplot(kind="strip", data=df_melted, x="Patient", y="Value", col="Feature", sharey=False)
plt.tight_layout()