Is there a way to hide nulls in an Altair chart?

Question

I don't want the null values to show on my Altair chart. Is there any way to do this without removing them from my data before charting them?

I have a data frame:

           d7_rent_own d8a_moving
0              Rent        NaN
1               Own         No
2               Own         No
3              Rent         No
4               Own         No
5              Rent         No
6              Rent        Yes
7               Own         No
8               Own         No
9               Own         No
10              Own         No
11              Own        Yes
12              Own         No
13             Rent        Yes
14             Rent         No
15              Own         No
16              Own         No
17              Own         No
18              Own         No
19              Own         No
20              Own         No
21             Rent        Yes
22              Own         No
23              Own         No
24              Own         No
25  No Answer Given        NaN
26             Rent        NaN
27             Rent         No
28              Own         No
29  No Answer Given        NaN
30              Own        NaN
31             Rent        Yes
32  No Answer Given         No
33             Rent        Yes
34              Own         No
35              Own        Yes
36             Rent         No
37             Rent         No
38              Own        NaN
39              Own         No
40             Rent        NaN
41             Rent        Yes
42             Rent        Yes
43             Rent         No
44              Own         No
45              Own         No
46             Rent         No
47             Rent        Yes
48             Rent         No
49             Rent         No

And I'm making a bar chart:

alt.Chart(df_ex).mark_bar().encode(
  alt.X('d7_rent_own:N', title = 'Housing Status'),
  alt.Y('count():Q', title = 'Number of Responses'),
  color = alt.Color('d8a_moving:N', legend=alt.Legend(title='Do you plan to move away?')))

But it shows up like this, with nulls:

Answer 1

You could drop these using pandas dropna on the dataframe as you are passing it to Altair. This will avoid removing them from the dataframe stored in the variable df_ex.

alt.Chart(df_ex.dropna()).mark_bar().encode(
  alt.X('d7_rent_own:N', title = 'Housing Status'),
  alt.Y('count():Q', title = 'Number of Responses'),
  color = alt.Color('d8a_moving:N', legend=alt.Legend(title='Do you plan to move away?')))

You could also use a transform filter in Altair for the same result:

alt.Chart(df_ex).mark_bar().encode(
  alt.X('d7_rent_own:N', title = 'Housing Status'),
  alt.Y('count():Q', title = 'Number of Responses'),
  color = alt.Color('d8a_moving:N', legend=alt.Legend(title='Do you plan to move away?'))
).transform_filter(
    'isValid(datum.d8a_moving)'
)