ENH generalize plot_ica_properties to channels

[jona-sassenhagen]

One EEGALB function I miss and find myself reproducing manually a lot is pop_prop.
These give a summary of a sensor or component that is helpful especially for identifying artefact types.
(They also work with raw files, segmenting the raw stream into 2-second epochs and showing you were gross artefacts are located in the data.)

plot_ica_properties should be generalized to channels. @mmagnuski

Maybe if others also find these interesting, we could implement them.

[jona-sassenhagen]

For channels rather than ICs, the sensor position is highlighted instead of showing the mixing weights.

[jona-sassenhagen]

If nobody is interested in this, I'll close.

[agramfort]

at least make a gist of your code so we know where to start to implement this.

[jona-sassenhagen]

closed by #3275

[mmagnuski]

Although the logic of OR (in the issue title) is fulfilled I think you can leave this open until channel properties are ready. :)

[jona-sassenhagen]

@agramfort I would like to also allow this for channels (like EEGLAB has). It's good to quickly see e.g. what kind of artefacts a channel has (just 2 or 3, or more). Ok?

mne.viz.plot_channel_properties(inst, ...)