altair grouped bar chart side-by-side instead of single chart

Question

I tried creating the chart however it ends up showing two charts side-by-side. How do I get a single chart instead like the one shown in altair grouped bar chart example.

df = pd.DataFrame({
    'strike': [200,225,250,275,300,325,350,200,225,250,275,300,325,350],
    'opttype': ['ce','ce','ce','ce','ce','ce','ce','pe','pe','pe','pe','pe','pe','pe' ],
    'oi': [100,150,500,800,450,200,77,50,500,210,300,150,60,17]
    
})
c = alt.Chart(df).mark_bar().encode(
    x='strike:N',
    y='oi:Q',
    color='opttype:N',
    column='opttype:N'
)

Output

I tried adding axis as specified in this link, however I could not get the combined graph. May be I am missing something here.

Also I wonder how the official documentation example works correcly without the "axis" as well.

jakevdp · Accepted Answer · 2021-11-19T01:33:36.540

1

The current best way to do this is to switch your x and column encodings, then make the width of each panel narrower:

c = alt.Chart(df).mark_bar().encode(
    x='opttype:N',
    y='oi:Q',
    color='opttype:N',
    column='strike:N'
).properties(width=50)

Then, if you wish, you can adjust axis & heading labels as in the linked examples. For example, you can move the header labels to the bottom like this:

c = alt.Chart(df).mark_bar().encode(
    x=alt.X('opttype:N', axis=None),
    y='oi:Q',
    color='opttype:N',
    column=alt.Column('strike:N', header=alt.Header(orient='bottom'))
).properties(width=50)

This is admittedly an imperfect solution. Vega-Lite is in the process of adding an offset encoding that will make this possible without having to use multiple column facets, but that is not available yet.

edited Nov 19 '21 at 01:33

answered Nov 18 '21 at 17:57

jakevdp

77,104
11
125
160

Thanks @jakevdp, this works. The offset would make this nicer. I'll play around to remove the labels (ce/pe) at the bottom. Also wondering if there is a way to get the Strike values at the bottom instead of the current position (at the top). My mind is trained to view the values at the bottom for X-axis. – Neo Nov 18 '21 at 18:47
Yes, the answer you linked to in your question shows how to move the heading labels to the bottom. – jakevdp Nov 18 '21 at 21:35
The example linked uses **"axis="** on the Column type which I do not see supported in the [latest docs](https://altair-viz.github.io/user_guide/generated/channels/altair.Column.html). The "spacing" option helped reduce the gap between the bars. – Neo Nov 19 '21 at 00:28
1

OK, see my edited answer – jakevdp Nov 19 '21 at 01:33

score 1 · Answer 2 · answered Nov 18 '21 at 17:57

1

The x-axis and color variable should be the same:

df = pd.DataFrame({
    'strike': [200,225,250,275,300,325,350,200,225,250,275,300,325,350],
    'opttype': ['ce','ce','ce','ce','ce','ce','ce','pe','pe','pe','pe','pe','pe','pe' ],
    'oi': [100,150,500,800,450,200,77,50,500,210,300,150,60,17]
})

alt.Chart(df).mark_bar().encode(
    x='opttype:N',
    y='oi:Q',
    color='opttype:N',
    column='strike:N'
)

answered Nov 18 '21 at 17:57

joelostblom

43,590
17
150
159

thanks @joelosblom for your answer. I just chose the previous as the accepted answer as that came in a few seconds earlier. – Neo Nov 18 '21 at 18:49

altair grouped bar chart side-by-side instead of single chart

2 Answers2