Forecast Plot for multiple models

[ngupta23]

Could you not have something like this pretty soon plot_model([model1, model2, model3]). I think that should be coming fairly quickly.

OR even

best_models = compare_models(n_select=3)
plot_models(best_models)

Originally posted by @ngupta23 in #1565 (comment)

[ejohnson-amerilife]

a workaround could be a for-loop over a list of models e.g.

for m in [model1, model2, model3]:
    plot_model(m)

[ngupta23]

for loops will plot separate plots for each model. What we want is one plot with overlays for multiple models so it makes it easy to compare visually.

Of course that does not preclude anyone from using the for loop. That could still be done as is today.

[ngupta23]

Expected behavior

# model1 supports Prediction Intervals (PI)
# model2 does not support PI
plot_model(model1)           # plots forecast with PI
plot_model(model2)           # plots forecast without PI
plot_model([model1, model2]) # plots both forecasts only (no PI)