Reputation: 1905

Seaborn columnwise violinplot

I have a dict d which lists numbers' occurences:

{'item1': [42, 1, 2, 3, 42, 2, 1, 1, 1, 1, 1],
 'item2': [2, 5],
 'item3': [5, 1, 7, 2, 7, 1, 42, 2, 9]}

Which I then convert to a DataFrame counting these occurences:

df = pd.DataFrame.from_dict({k: dict(Counter(v)) for k, v in d.items()})

    item1  item2  item3
42    2.0    NaN    1.0
1     6.0    NaN    2.0
2     2.0    1.0    2.0
3     1.0    NaN    NaN
5     NaN    1.0    1.0
7     NaN    NaN    2.0
9     NaN    NaN    1.0

How can I plot this or some other DataFrame that was derived from d using seaborn.violinplot, so that each column in the dataframe represents a violin in the plot based on the data provided by each columns values and their respective indices?

I have tried multiple combinations of which I believe this intuitively comes the closest, but unfortunately still fails:

sns.violinplot(x=df.keys(), y=df.index, data=df)

Upvotes: 1

Answers (3)

Nilesh Ingle

Reputation: 1883

Code below creates the plot:

# Import libraries
import pandas as pd
import collections
import seaborn as sns

# Create dictionary
d = {'item1': [42, 1, 2, 3, 42, 2, 1, 1, 1, 1, 1],
 'item2': [2, 5],
 'item3': [5, 1, 7, 2, 7, 1, 42, 2, 9]}

# Create DataFrame and melt
df = pd.DataFrame.from_dict({k: dict(collections.Counter(v)) for k, v in d.items()})
df = df.melt()

# Plot
sns.violinplot(x="variable", y="value", data=df)

Upvotes: 2

Diziet Asahi

Reputation: 40747

Simply this is enough

from collections import Counter
d = {'item1': [42, 1, 2, 3, 42, 2, 1, 1, 1, 1, 1],
     'item2': [2, 5],
     'item3': [5, 1, 7, 2, 7, 1, 42, 2, 9]}
df = pd.DataFrame.from_dict({k: dict(Counter(v)) for k, v in d.items()})

sns.violinplot(data=df)

Upvotes: 2

jezrael

Reputation: 863611

Pass DataFrame to seaborn.violinplot, so each Series (column) is plotted separately:

sns.violinplot(data=df)

Upvotes: 2

Seaborn columnwise violinplot

Answers (3)

Related Questions