`plotly.express.scatter_geo`: misleading argument type for the `size` argument

[stoufa]

The documentation of the function plotly.express.scatter_geo [1] states that the size argument can be set to an integer.
When you do that, you'll get an error similar to the following:

Traceback (most recent call last):
  File "/path/to/map.py", line 11, in <module>
    fig_map = px.scatter_geo(
  File "/path/to/miniconda3/lib/python3.8/site-packages/plotly/express/_chart_types.py", line 1031, in scatter_geo
    return make_figure(
  File "/path/to/miniconda3/lib/python3.8/site-packages/plotly/express/_core.py", line 1861, in make_figure
    args = build_dataframe(args, constructor)
  File "/path/to/miniconda3/lib/python3.8/site-packages/plotly/express/_core.py", line 1377, in build_dataframe
    df_output, wide_id_vars = process_args_into_dataframe(
  File "/path/to/miniconda3/lib/python3.8/site-packages/plotly/express/_core.py", line 1183, in process_args_into_dataframe
    raise ValueError(err_msg)
ValueError: Value of 'size' is not the name of a column in 'data_frame'. Expected one of ['col_name_1', 'col_name_2', 'col_name_3'] but received: 20

After some research, [2] [3] I found out that the correct way of doing this is via the method fig.update_traces(marker=dict(size=20))
So, we need to update the documentation to avoid obsolete and misleading argument types

[nicolaskruchten]

The size argument can indeed be set to an integer to refer to data frame columns whose indices are integers.

[stoufa]

Thanks for the swift reply, I thought that the integer would be the size of the marker not the column index in the data frame.
This detail wasn't mentioned on the documentation page. We definitely should add this information and send readers to the right page if they want to make the marker size the same for all points.

[nicolaskruchten]

I'll see what I can think of to make this clearer.

Note that you can get the effect you're looking for with size=px.Constant(<size in pixels>).

Also, this same "pointer or value" ambiguity happens with arguments like color, in that color="red" will not yield all-red points, it will bind color to a column called "red", but you can get all-red points with color=px.Constant("red").

[stoufa]

Perfect! I didn't know about this px.Constant trick either.
By adding these tiny details, we won't have any confusion anymore.
Thanks again @nicolaskruchten.

[nicolaskruchten]

Yeah, I should really document both the "integers are interpreted as col names" and "px.Constant is a valid input" in the docstrings. Thanks for the nudge :)

[stoufa]

I have a tiny question about the same parameter: The size of the marker appears with the hover_data, is there a way to hide it? — It looks a bit confusing especially when it's the same for all the markers.

[nicolaskruchten]

Yes: https://plotly.com/python/hover-text-and-formatting/#disabling-or-customizing-hover-of-columns-in-plotly-express

[stoufa]

Thank you @nicolaskruchten.
For reference, using fig.update_traces(marker=dict(size=20)) instead of size=px.Constant(20) is the fastest way to do that.

[gvwilson]

Hi - we are trying to tidy up the stale issues and PRs in Plotly's public repositories so that we can focus on things that are still important to our community. Since this one has been sitting for several years, I'm going to close it; if it is still a concern, please add a comment letting us know what recent version of our software you've checked it with so that I can reopen it and add it to our backlog. Thanks for your help - @gvwilson