Annotated heatmap with multiple color schemes from two different columns of dataframe-CodePudding

I have a dataframe in the following format:

Percent category  MolA  MolB
  20      1     02K@sc  81G@bs
  15      0     04F@sc  79U@bs
  16      2     04F@sc  79U@sg
  18      1     04F@sc  80A@bs
  4       0     04F@sc  80A@ph

I want to plot the MolA column against MolB according to percent value. For coloring each category should be in different color for example 0 -->Blue, 1 --> Red, 2-->Green and then the hue/shades of these color will be according to percentage.

I tried using the sns.colorpalette("YlGnBu", 30) but this is not exactly what I want and this seems to be not very clear for distinguishing between two different categories.

Is it possible to plot the dataframe in such way?

CodePudding user response：

It is possible to leverage the behavior of heatmap automatically excluding null values from the map by iterating over each category, nulling out Percent not in that category and making a heatmap for the resulting pivot table. If you loop that you may be able to get what you want.

import pandas as pd
df = pd.DataFrame({'Percent': [20, 15, 16, 18, 4],
 'category': [1, 0, 2, 1, 0],
 'MolA': ['02K@sc', '04F@sc', '04F@sc', '04F@sc', '04F@sc'],
 'MolB': ['81G@bs', '79U@bs', '79U@sg', '80A@bs', '80A@ph']})


palettes = ['Blues','Greens','Oranges']

for c, p in zip(df.category.unique(), palettes):
    t = df.copy()
    t['Percent'] = np.where(t['category'].ne(c), np.nan, df['Percent'])
    sns.heatmap(t.pivot('MolA','MolB','Percent'), cbar=False, cmap=p, annot=True)