add ability to plot sample output_type data

[nickreich]

Currently, a limitation of plot_step_ahead_model_data() is that it only accepts median and quantile output_types.

We should support the plotting of sample output_type data by inferring quantiles from it. Currently, a user would have to manually transform the data to quantiles. Possible approach: add sample output_type to the list of valid types and make the appropriate transformation to quantile output_type data.

Acceptance criteria:

• a user passes sample output_type data to plot_step_ahead_model_output() and gets a plot returned, with intervals if specified

[LucieContamin]

Would it make sense to have a choice to either:

• calculate the quantiles and return a plot with intervals as specified
• return a "spaghetti plot" of the samples

[elray1]

i like that, and would add "or both"

[LucieContamin]

"both", meaning on the same plot? It might get a little crowded/complex to read but we can try some combination with maybe a restriction on the number of intervals in this case. We can run some tests.

[elray1]

yeah, I was thinking both on the same plot. i've seen this done before and found it helpful, but agree with your comment that it could get crowded. thinking about it now, i think the examples I saw showed only a few of the samples.

one note about showing the samples is that it is not guaranteed that the samples at different horizons can be connected... we'd need to check that they were actually trajectories. (you probably thought of that already, just wanted to get it down). or maybe use geom_line if they are trajectories and geom_point otherwise?

[LucieContamin]

Yes, I was thinking the same for the samples, we can use point or line depending if they are trajectories or not. I like it and should not be difficult to implement

[elray1]

this point relates to hubverse-org/hubData#6 as well