matplotlib - Seaborn chart colors are different than those specified by palette

Question

Why are seaborn chart colors different from the colors specified by the palette?

The following two charts show the difference between the colors as they appear on a bar chart, and the colors as they appear in the palette plot. You can see if yo ulook carefully, that the colors on the bar chart are slightly less bright/saturated.

Why are these different, and how can I get the bar chart to have the exact same colors as the ones specified in the palette?

import seaborn as sns
sns.set(style="white")
titanic = sns.load_dataset("titanic")

colors = ["windows blue", "amber", "greyish", "faded green", "dusty 
purple"]

ax = sns.countplot(x="class", data=titanic, 
palette=sns.xkcd_palette(colors))
sns.palplot(sns.xkcd_palette(colors))

Bar chart

Palette plot

Answer 1

Many seaborn plotting commands have an argument saturation, whose default value is 0.75. It sets the saturation (S) of the colors in the HSL colorspace (ranging from 0 to 1) to the given value.

Setting this parameter to 1 in the countplot will give you the same colors in both plots.

ax = sns.countplot(x="class", data=titanic, palette=sns.xkcd_palette(colors), saturation=1)
sns.palplot(sns.xkcd_palette(colors))

The reason for this default desaturation is that many people consider a plot with less contrast to be more appealing. That is also why the default background in seaborn is not white but some bluish gray. After all, this is of course a question of taste.