Altair: Grouped bar chart: Remove X values for which Y values are null

[selmaca]

Hi,

In a grouped bar chart, is it possible to remove X values for which there are no Y values in a certain group?

I am plotting the number of prosumers and consumers on different buses in three different distribution systems. Each distribution system (a group on a bar chart) has different buses, and I don't want to show the ones that are not there. Is there a property that would allow me to mask the X values for which Y is 0 or Null/NaN? ..or should I use a completely different plotting?

I have a very basic chart code:

alt.Chart(merged_df).mark_bar().encode(

y=alt.Y('Number of agents:Q'),

x=alt.X('Buses:N'), 

column=alt.Column('DSs:N') ,

color=alt.Color('Type:N'))

Thanks!

[jakevdp]

You can do this by using a facet chart (rather than a facet encoding) and setting the x-scales to independent:

alt.Chart(merged_df).mark_bar().encode(
  y=alt.Y('Number of agents:Q'),
  x=alt.X('Buses:N'), 
  color=alt.Color('Type:N'),
).facet(
  column=alt.Column('DSs:N'),
).resolve_scale(
  x='independent'
)

[selmaca]

Hi Jake,

Thanks, but that code gives me exactly the same chart as in the original post.

Any idea why setting x='independent' seems to do nothing?

[jakevdp]

If you could include a reproducible example of what you tried, I may be able to answer your question. Otherwise, it's just guesswork :smile:

[jakevdp]

Here is what I tried. As you can see, empty columns in each chart are hidden:

import altair as alt
import numpy as np
import pandas as pd

np.random.seed(0)

merged_df = pd.DataFrame({
    'Number of agents': np.random.randint(1, 4, 20),
    'Buses': np.random.randint(1, 17, 20),
    'DSs': np.random.randint(0, 3, 20),
    'Type': np.random.choice(['Consumers', 'Prosumers'], 20)
})

alt.Chart(merged_df).mark_bar().encode(
  y=alt.Y('Number of agents:Q'),
  x=alt.X('Buses:O'), 
  color=alt.Color('Type:N'),
).facet(
  column=alt.Column('DSs:N'),
).resolve_scale(
  x='independent'
)

[selmaca]

Hi Jake,

Sorry! I just made a test dataframe with the same data and it worked perfectly!

I realized that the masking of the original dataframe was not working properly, but now, everything is fixed.

thanks for the help!

[jakevdp]

Great!