CODEWITHSUNDEEP
pythonaltair
Omega
Omega

Reputation: 871

Altair selection with different colours

I started playing around with Altair today and have a question about selections. I have created 3 bar charts with different colours as shown below. Now whenever I click on one bar, I'd like the equivalent to be highlighted in the other two bar charts. This works well if I don't use any colour and it defaults to blue, but I can't figure out how to keep my colours in the selection instead of having the same colour for all 3 charts.

enter image description here

df = pd.DataFrame({"Name":["Bulbasaur", "Charizard", "Mewtwo"], 
                  "HP":[45, 80, 100], 
                  "Attack":[30, 50, 60], 
                  "Defense":[40, 38, 42], 
                  "Type":["Grass", "Fire", "Psychic"]}) 


selection = alt.selection_single(fields=["Type 1"])

# This doesn't do much as I don't have a column named #73a1eb. But it's the colour code I'd like to use.
color1 = alt.condition(selection,
                  alt.Color("#73a1eb:N", legend=None), 
                  alt.value('lightgray'))

color2 = alt.condition(selection,
                  alt.Color("#73b9c7:N", legend=None),
                  alt.value('lightgray'))

color3 = alt.condition(selection,
                  alt.Color("#d7abf5:N", legend=None),
                  alt.value('lightgray'))

# The mark bar colour is overriden by color in encode

chart1 = alt.Chart(df, title="Average HP by Type")
            .mark_bar(color="#73a1eb", size = 12)
            .encode(x = 'mean(HP):Q',
                    y = alt.Y('Type 1:N', sort='-x'), 
                    tooltip=["Type 1", "mean(HP):Q"], 
                    color=color1)
            .properties(height=320,
                        width=300)
            .add_selection(selection)

chart2 = alt.Chart(df, title="Average Attack by Type")
            .mark_bar(color="#73a1eb", size = 12)
            .encode(x = 'mean(Attack):Q',
                    y = alt.Y('Type 1:N', sort='-x'), 
                    tooltip=["Type 1", "mean(HP):Q"], 
                    color=color1)
            .properties(height=320,
                        width=300)
            .add_selection(selection)

chart2 = alt.Chart(df, title="Average Defense by Type")
            .mark_bar(color="#73a1eb", size = 12)
            .encode(x = 'mean(Defense):Q',
                    y = alt.Y('Type 1:N', sort='-x'), 
                    tooltip=["Type 1", "mean(HP):Q"], 
                    color=color1)
            .properties(height=320,
                        width=300)
            .add_selection(selection)

alt.hconcat(chart1, chart2, chart3).configure_axis(grid=False).configure_axisBottom().resolve_scale(x = 'shared')

Upvotes: 1

Views: 2961

Answers (1)

jakevdp
jakevdp

Reputation: 86310

When you write alt.Color("#73a1eb:N"), it means you want the color to be encoded according to a column named "#73a1eb" which has a nominal ("N") type.

It appears that you want to specify a color value rather than a color encoding, in which case you can write alt.value("#73a1eb"). So your condition would look like this:

color1 = alt.condition(selection,
                       alt.value("#73a1eb"),
                       alt.value('lightgray'))

Upvotes: 4

Related Questions