Improve GAM output data visualization

[damianooldoni]

When we correct by using cobs as covariate, the classic output (via plot_ribbon_em function) doesn't give the best visualization. See example below:

This is due to the effect of the covariate. Is there a way to snow the effects of covariates in a plot?
Via draw(deriv1) I can see the effects of covariates on first derivative. It would be nice to see similar plots for the data itself. Do you know how to do it? Thanks!

[damianooldoni]

To compare, without covariate:

So, the actual way of plotting is good for basic GAM but less intuitive if we correct based on class data.