CODEWITHSUNDEEP
pythonplotnine
Jameson
Jameson

Reputation: 187

Changing color of each bar for geom_bar in plotnine

I'm making some bar charts in plotnine of counts on a single variable (based on survey responses).

I'm currently using this code:

(ggplot(df, aes ('age_cat3')) +
geom_bar(color='#F6F6F6', fill='#0C475B') +
theme_minimal() +
xlab("") +
ggtitle('Age distribution in total sample'))

And this produces:

enter image description here

I would like each bar to have a different color. I know I could do this by adding [fill= "age_cat3"], so:

(ggplot(df, aes ('age_cat3', fill='age_cat3')) +
geom_bar() +
theme_minimal() +
xlab("") +
ggtitle('Age distribution in total sample'))

And this produces:

enter image description here

There are two issues with this solution:

  1. There in now a legend, which I can't seem to remove with "scale_color_discrete(guide=False)"

  2. I can't seem to control the color or color theme.

Any advice on removing the legend would be great, but ideally, some way of changing the color of the bars without having to use "fill" would be great.

(also, sorry the images came out so big. I'm not quite sure how to attach small images)

reproducible data:

df = pd.DataFrame({'id': [1, 2, 3, 4, 5, 6, 7, 8 , 9, 10],
                   'age_cat3': [15-24, 15-24, 25-34, 35-44, 35-44, 35-44, 35-44, 45-59, 45-59, 60+]}

Upvotes: 1

Views: 4033

Answers (2)

Guilherme Parreira
Guilherme Parreira

Reputation: 1021

Another option to remove the legend is:

(ggplot(...)
...
+ theme(legend_position = element_blank())

or place it in another position:

(ggplot(...)
...
+ theme(legend_position = 'top')

The scale_fill_manual is the best and maybe the only way to include color values manually.

Upvotes: 1

has2k1
has2k1

Reputation: 2375

You can turn off the legend with geom/layer parameter show_legend=False and you can use a manual scale to set the fill colours.

(ggplot(...)
...
+ geom_bar(show_legend=False)
+ scale_fill_manual(values=['red', 'blue', 'green', 'cyan', 'violet', '#AABBCC'])
)

Upvotes: 2

Related Questions