Feature planning.

[liamzebedee]

• Move back and forth by month in progress report.
• Filter progress report by events, ie. days with "Drinking".
• Links to individual days.
• Creating reports - composing different views of data. ie. basal tests, sickness events.

[liamzebedee]

Requirements analysis.

• Move back and forth by month in progress report.

  • Why? To quickly view monthly patterns. (ie. seasonal/lifestyle changes).
  • How often do we move back, 6 months? Once or twice.
  • How often do we move back, 6 months? Not so often.
  • How often do we move back, 12 months? Not so often.
  • Would it make sense to toggle between 30 day view and monthly view? Monthly should be enough.
  • What is more logical? Monthly or 30 days rolling? Would it be useful to do both?
  • Should the PGS/HBA1c estimate change contextually? I would think so. Is it useful to have a longer-term value here? Not necessarily.

• Filter progress report by events.

  • Why? To quickly glance at different days with the same event to identify patterns. ie. sickness. What is the best view for this? Yearly for sickness, monthly for exercise.
  • What types of events do we have? These are based on the user. Dynamic checklist.

• Links to individual days.

  • Why? To support external outside-of-app organisation of data.
  • Who is using these links? The user. Do we need permissions/ACL? Yes. What if we created links for other people to collaborate with? Building a platform for diabetics to share data and spot patterns? Different permission scheme would apply.

• Creating reports.

  • Why? To generate insights off data. ie. what happens when I'm sick/stressed, what happens when I eat pizza.
  • Are they based on days?
    • Sickness is day-based probably.
    • Exercise is roughly hour-based.
    • Caveat: effects may be present beyond an inter-day boundary - ie. exercise at 6pm, reduced insulin requirements into the next morning.
    • A: they are based on time slices, which could cross days.
  • Is the data useful overlaid, or interpretable? It will likely depend. We could let user download a CSV for their own analysis. Probably easier if it's just a collection of days.
  • Do the reports have permissions? Link-only.
  • What is the utility of the reports? Filtering the data down to generate a view of the same event (ie. exercise, drinking).
  • Are the reports static / one-time? Perhaps. You could imagine generating a report for sickness, and sharing that with other diabetics. Then when you get sick again, would that change the original story of the report? Perhaps. It would be interesting to see how other diabetics have coped with the stress induced by the COVID-19 pandemic.

[liamzebedee]

Prioritisation.

Creating reports most useful for value.
A "manual" report could be made using:
  - link to day/slice
  - download CSV
How do we get the links to different days?
 - find the days manually
 - remember the dates
 - search for events
The easiest is to search for events, and then select via preview + checkbox.

Search for days
Select days
Search for events
Select events
Select other bits
Generate report

What's the manual equivalent?

Find the days. <------- ! Block: how do we find the days with events? Easiest - navigate and click in progress report. Does this include event markers? No. For this, we need some filtering.
Download JPEG's.
Insert in Gdoc.

What if there are multiple of the same event in a day? 
  - select via checkbox?

Moving back and forth month-by-month is good for spotting patterns.

Technical analysis.

Links for days.
  Integrate a routing solution. eg. react-router.
  Build new IndividualView component.
  Build URL's.
  Add UI element to progress report -> View day in logbook.
Download JPEG/CSV.
  Render