Difficulties importing data into Seaborn Boxplot

Question

EDIT: A solution was provided by user #kgoettler below. The problem arise from Seaborn Boxplot requiring data to be organized by variable in the x-axis and values in the y-axis. The script below reorganizes the data into a form compatible with Seaborn Boxplot.

ORIGINAL QUESTION: My goal is to generate a Box plot using data from an CSV file. I would like to use the Python visualization library Seaborn. The data is organized with a common index (Object) and headers for each column.

I have difficulties importing this data into a Boxplot using the format

seaborn.boxplot(x="variable", y="value")

Using Pandas own boxplot this is not a problem since I simply specify what columns to use based on headers using the following format

boxplot = data.boxplot(column=['header1', 'header2', 'header3'])

I would also prefer not to have to specify each individual column by header, but rater select all columns in the file automatically.

All feedback and input is greatly appreciated!

kgoettler kgoettler · Accepted Answer · 2020-04-14T19:23:23

Something like this should work:

import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
sns.set(style='whitegrid')

csv_file = '/path/to/file.csv'
df = pd.read_csv(csv_file)
df = (df
        .set_index(['Object'])          # Set 'Object' column to index
        .rename_axis("Metric", axis=1)  # Rename the column axis "Metric"
        .stack()                        # Stack the columns into the index
        .rename('Score')                # Rename the remaining column 'Score'
        .reset_index()                  # Reset the index
    )

This should give you a DataFrame that looks like:

   Object                       Metric     Score
0     MT1  B1A1 Average Splaying Score  0.426824
1     MT1  B1A2 Average Splaying Score  0.431351
2     MT1  B1A3 Average Splaying Score  1.941473
3     MT2  B1A1 Average Splaying Score -0.021672
4     MT2  B1A2 Average Splaying Score  3.357387

Then to plot, all you have to do is:

fig, ax = plt.subplots(figsize=(10,6))
ax = sns.boxplot(x='Metric', y='Score', data=df, ax=ax)
ax.set_xlabel('')

Example Plot

Difficulties importing data into Seaborn Boxplot

1 Answers