Add vignette on modelling temporally aggregated data (i.e., obs_opts(na = "accumulate") )

[jamesmbaazam]

Thanks a lot all - I've tested it out and it's working well now! I've included the updated figure below on the same data for reference, in case helpful!

@jamesmbaazam regarding extra documentation, I think that at some point it could be useful to add a short, dedicated section on applying to weekly data. The current documentation mentions the rw estimator for weekly, but I think having a section with an example as you do for daily could be beneficial to others as we work a lot with weekly aggregated data. The other thing would probably be then to be able to visualize on weekly scale too, but as it's possible for end users to extract the estimates and create visuals I don't think it's really necessary.

Thanks again for your help and for a great package!

Originally posted by @martinamcm in #767 (comment)

[jamesmbaazam]

By "surface", I mean we need a vignette walkthrough. My argument for this need is here #774 (comment).

[seabbs]

Agree. Perhaps edit the title and add a plan/outline?

[jamesmbaazam]

Outline of vignette

• Discuss how EpiNow2 treats/prepares input data (into complete daily data)
• Motivate for data being reported at different types of intervals;
  • daily,
  • irregular intervals (every few days), and
  • regular but at an aggregated scale (weekly, monthly)
• Walk-throughs of how EpiNow2 supports observations reported at non-daily intervals:
  • Treating missing dates as explicitly missing (obs_opts(na = "missing"))
  • Treating missing dates as accumulated (obs_opts(na = "accumulate"))

[jamesmbaazam]

This vignette will benefit from the merge of #774 as it will showcase the message thrown when the default is used.

[seabbs]

We might want to wait on this until we have some sort of clarity on the approach that is going to be used to deal with #547?

[jamesmbaazam]

Agreed. Alternatively, if written with current functionalities, it can be revised to use whatever is changed after #547.

[seabbs]

Sure but if any change is imminent that might be more work? On the other hand if we still need to discuss what #547 looks like it could help inform that discussion.

[jamesmbaazam]

Agrees. Let's hold this off for now. The question about treatment of missingness has been coming up a lot lately.