Conditionally color pandas dataframe by column

Question

I am working on a jupyter notebook right now and I am looking for a way to conditionally color each cell in a pandas dataframe according to its relative value within the column (or alternatively row).

The final output should be a pandas dataframe.
Conceptually it would be like creating a heatmap where the shading is defined independently for each column and is based on the max and min of the column itself.

I have had a look at this and this but in both they create actual plot as output instead of coloring the dataframe cells.

right, I am sorry. I am working on a jupyter notebook right now =) I'll update the question — CAPSLOCK, Nov 21 '19 at 12:42
How do you display your dataframe? Have you looked at this: https://pandas.pydata.org/pandas-docs/stable/user_guide/style.html ? — Karl Anka, Nov 21 '19 at 12:43

score 6 · Answer 1 · answered Nov 21 '19 at 12:57

6

You can find more options here: Pandas DataFrame Styling: Builtin Styles

import seaborn as sns
cm = sns.light_palette("green", as_cmap=True)

df.style.background_gradient(cmap=cm, axis=0) # explicitly applying column-wise

The output will look like:

answered Nov 21 '19 at 12:57

CypherX

7,019
3
25
37

score 2 · Accepted Answer · edited Nov 21 '19 at 12:49

2

You can use the style method. The output looks like a dataframe, but is not.
If you want each column to have a color gradient corresponding to the values of the column:

df.style.background_gradient()

To apply the style row-wise, use the additional parameter axis=1.

edited Nov 21 '19 at 12:49

CAPSLOCK

6,243
3
33
56

answered Nov 21 '19 at 12:44

Horace

1,024
7
12

Conditionally color pandas dataframe by column

2 Answers2