Text color as function of non-numeric column value

[dafriedman97]

I have a DataFrame with a non-numeric column and I want to color text as a function of that column's values. I also have a specific color mapping in mind. So say I have the following table

df = pl.DataFrame({
    'city': ['New York', 'Boston', 'Chicago', 'Seattle'],
    'region': ['east', 'east', 'central', 'west']
})

┌──────────┬─────────┐
│ city     ┆ region  │
│ ---      ┆ ---     │
│ str      ┆ str     │
╞══════════╪═════════╡
│ New York ┆ east    │
│ Boston   ┆ east    │
│ Chicago  ┆ central │
│ Seattle  ┆ west    │
└──────────┴─────────┘

and a color mapping like {'east': 'green', 'central': 'blue', 'west': 'red'}. How can I format the city name with its mapping color?

[dafriedman97]

Also curious how I could make the background color (for a column or the entire row) depend on the value. Thanks!

[jrycw]

Here’s my approach:

import polars as pl
from great_tables import GT, style, loc

df = pl.DataFrame(
    {
        "city": ["New York", "Boston", "Chicago", "Seattle"],
        "region": ["east", "east", "central", "west"],
    }
)

gt = GT(df)

color_map = (
    pl.when(pl.col("region").eq(pl.lit("east")))
    .then(pl.lit("green"))
    .when(pl.col("region").eq(pl.lit("central")))
    .then(pl.lit("blue"))
    .when(pl.col("region").eq(pl.lit("west")))
    .then(pl.lit("red"))
)

# Format the city name using its corresponding mapping color.  
(gt.tab_style(style=style.text(color=color_map), locations=loc.body("city")))

# Format the background color using its mapping color by specifying a column (or multiple columns). 
(gt.tab_style(style=style.fill(color=color_map), locations=loc.body("city")))

# Format the background color using its mapping color by specifying a column (or multiple columns) 
# and/or a row (or multiple rows).
(gt.tab_style(style=style.fill(color=color_map), locations=loc.body("city", rows=[0, 2])))

Hope this helps!

Reference: Basic Styling in Great Tables.

[dafriedman97]

Helps a ton. Thank you!!

[dafriedman97]

Would be nice if there was a way that used a dictionary instead of writing out the when/then logic though (in reality I have many more fields/colors)

[jrycw]

No worries! The pl.when().then().otherwise() approach works well for numerical columns when applying logic based on calculations. However, if the column is more categorical, you might try a simple replacement using a predefined color mapping like:

color_mapping = {
    "east": "green",
    "central": "blue",
    "west": "red",
}


(
    gt.tab_style(
        style=style.text(color=pl.col("region").replace(color_mapping)),
        locations=loc.body("city"),
    )
)

[jrycw]

As this is a practical question, I’ve decided to summarize what I know in a blog post, which you can find here.

[dafriedman97]

I appreciate that a ton! I didn't know about using pl.col() within style.text or about from_column(). Thanks so much!