如何在一个 plot 和 seaborn 上绘制数据网格的群图 plot

Question

I having been trying for ages to plot this data the way I want on seaborn, from a pandas dataframe. Any advice would be appreciated.

The Data is in a dataframe and looks like this, with 12 months, and 11 columns plus a month column:

  Month   FSRPX1M     VOX1M     XLB1M  ...     XLP1M     XLU1M     XLV1M     XLY1M
0   Jan  0.087643 -0.561428 -0.409286  ... -0.177143  0.232858  0.521428  0.457857
1   Feb  0.132429 -0.265715  0.470715  ...  0.067142 -0.479286 -0.177143  0.363571
2   Mar  0.152429  1.002142  0.437857  ...  0.588573  0.671428  0.055000  0.727856
3   Apr  0.150071  2.445000  1.331428  ...  0.551428  1.100715  0.790715  2.175714
4   May -0.089429 -0.115714  0.275000  ...  0.240000  0.113572  0.650716 -0.366429

What I want to do is plot a swarm plot, with Months along the X Axis, and the number along the Y axis, with 11 'dots' for each month, representing the 11 columns, with an associated a color coded key.

I have tried various things including [where df is the dataframe and val_list is a list of the columns I want as dots]:

sns.swarmplot(data = df, x=df['Month'], y = df[val_list])

and

sns.swarmplot(data = df, x=df['Month'], hue = df[val_list])

as well as trying to slice the dataframe, and make a plot, one column at a time using a for loop:

for e in val_list:
    sns.stripplot(data = df, x=df['Month'], y = df[e])

Any help and explanation of what I'm doing wrong would be great, thanks

The last one makes a plot, but all the dots are the same color and unlabelled

Answer 1

To work with hue , Seaborn needs the data in "long form" . Pandas' melt can be used to convert a "wide" dataframe to a "long" one. The Set3 colormap has 12 colors (the default tab10 only has 10).

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

elements = ['Hydrogen', 'Helium', 'Lithium', 'Beryllium', 'Boron', 'Carbon',
            'Nitrogen', 'Oxygen', 'Fluorine', 'Neon', 'Sodium']
df = pd.DataFrame(np.random.randn(12, len(elements)), columns=elements)
df['Month'] = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']

df_long = df.melt(id_vars='Month', var_name='Element', value_name='Value')

ax = sns.swarmplot(data=df_long, x='Month', y='Value', hue='Element', palette='Set3')
ax.legend(bbox_to_anchor=(1.01, 1.02), loc='upper left', title='Element')
plt.tight_layout()
plt.show()