how to position labels in a geom_bar?

[syd-doyen]

Hi Everyone,

I'm really happy to have found lets-plot. It has made visualising data pretty easy.

I have the following issue, I'm having a hard time understanding how to place labels in a geom_bar plot. Ideally, I want to have control over where in the bar they are drawn. Middle, outside the bar, etc.

But I can't move the labels. I feel that by using alignment characters ^ < > is only for the characters specified by the format not the graphic the text is bounded by.

This has also being a problem when creating a stacked geom_bar plot.

Below is my example

import pandas as pd
import lets_plot as lp
from lets_plot.mapping import as_discrete

lp.LetsPlot.setup_html()

df_female = pd.DataFrame(data={'Industry': ['Construction', 'Mining', 'National',
                                                 'Construction', 'Mining', 'National'],
                                    'Category': ['Female', 'Female', 'Female',
                                                 'Female Manager', 'Female Manager', 'Female Manager'],
                                    'Proportion': [26.0,  22.0, 51.2, 18.9, 22.9, 42.0]})

ordered_industry = as_discrete('Industry', order=-1)
p2 = (lp.ggplot(df_female, lp.aes(weight='Proportion', x=ordered_industry, fill='Category')) +
    lp.geom_bar(lp.aes(fill='Category'),
            position=lp.position_dodge(width=0.9),
            labels=lp.layer_labels().line('@..count..').format(field='@..count..',format="^.0f").size(15), alpha=0.7) +

    lp.labs(title=' (Proportion by industry, 2023)',
              x='', y='Female Proportion (%)', fill='') +
    lp.scale_fill_manual(values=['#FFC300', '#FFEA33'], name='',
                           breaks=['Female_p','Female_Manager_p'], labels=['% Women', '% Women managers']) +
    lp.theme_bw() + lp.theme(legend_position="top"))

Here is what I get

Any help is much appreciated.

[OLarionova-HORIS]

Hello,

Unfortunately, there is currently no control over the position of the bar annotations.
As an alternative, geom_text or geom_label layers could help.

[syd-doyen]

Thanks, I will try your suggestions

[OSuwaidi]

@syd-doyen hey, you might find this of help: #978 (comment)