Change scales for repeated Charts

Not all repeated columns / rows have the same scales. It would be useful not enable different for different columns / rows

I do not know what the API should look like since I am not very familiar with this code base. but I thought of this: make repeated able to take fully specified encoding to be modified

import altair as alt
from vega_datasets import data
iris = data.iris.url

chart = alt.Chart(iris).mark_point().encode(
    alt.repeat("column"),
    alt.Y(alt.repeat("row"), type='quantitative'),
    color='species:N'
).properties(
    width=200,
    height=200
).repeat(
    row=[
        alt.X('petalLength', type='quantitative', scale=alt.Scale(base=10, log=True)), 
        alt.X('petalWidth', type='quantitative', scale=alt.Scale(base=10, log=False))],
    column=['sepalLength', 'sepalWidth']
).resolve_scale().interactive()

The code above raises the error below

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
<ipython-input-13-573643beab36> in <module>
      6     alt.repeat("column"),
      7     alt.Y(alt.repeat("row"), type='quantitative'),
----> 8     color='species:N'
      9 ).properties(
     10     width=200,

C:\anaconda3\lib\site-packages\altair\vegalite\v3\api.py in encode(self, *args, **kwargs)
   1385     def encode(self, *args, **kwargs):
   1386         # Convert args to kwargs based on their types.
-> 1387         kwargs = utils.infer_encoding_types(args, kwargs, channels)
   1388 
   1389         # get a copy of the dict representation of the previous encoding

C:\anaconda3\lib\site-packages\altair\utils\core.py in infer_encoding_types(args, kwargs, channels)
    467         if encoding is None:
    468             raise NotImplementedError("positional of type {}"
--> 469                                         "".format(type_))
    470         if encoding in kwargs:
    471             raise ValueError("encoding {} specified twice.".format(encoding))

NotImplementedError: positional of type <class 'altair.vegalite.v3.schema.core.RepeatRef'>

[jakevdp]

Hi @crypdick – I don’t think there is a way to adjust individual scale parameters in a repeated chart. To do that, you’d have to manually concat the subcharts.

P.S. Good to see you again 😁

[crypdick]

@jakevdp One thing I haven’t been able to find documentation for is how to make some columns of a repeat chart log-scale. For example, how do I make the learning rate and L2 penalty x-axes log-scale (but not dropout)?

    chart = alt.Chart(data).mark_circle().encode(
        alt.X(alt.repeat("column"), type='quantitative',scale=alt.Scale(zero=False, padding=0.01)),
        alt.Y(alt.repeat("row"), type='quantitative', scale=alt.Scale(zero=False, padding=0.01)),
    ).repeat(
        column=['Learning Rate', 'L2 Penalty', 'Dropout'],
        row=['val/avg_precision' ]
    )

P.S. I just discovered Altair and it’s already my favorite Python viz package. P.P.S. you taught me how to use Numpy circa 2014 at UW P.P.P.S. Sorry for the necro post