plotting posterior responses distribution

[AdityaSavara]

PEUQSE should have an option to keep all the responses and then filter out the unlikely ones and then to plot them.

This would be similar to drawing from the posterior:
az.plot_ppc(az.from_pymc3(posterior_predictive=ppc, model=model_1));
https://docs.pymc.io/en/stable/pymc-examples/examples/diagnostics_and_criticism/posterior_predictive.html

In general, should probably start using ArviZ for plotting.
https://arviz-devs.github.io/arviz/user_guide/pymc3_refitting_xr_lik.html