Population - Site population, polygon population, and population measures
mathew-thomson opened this issue · 2 comments
It was brought up that population is an important aspect to capture, and it was initially suggested that this should be exclusively for 'Site' since it is a “slowly changing dimension” – seldomly changing, if ever unlike a real field measurement. As such, it could almost be treated as a fixed attribute, like zipCode. It’s also (by and large) an estimate, except for when referring to airplanes, cruises, etc.
Everyone seems to agree that population should be an attribute of a site, but we also think that population could also be a Polygon attribute (e.g., if a polygon represents a public health region, it doesn’t necessarily have a site, however, the population living in that health region is important and should be recorded somewhere).
We also think think that population should be a measure, as for some sites population changes day-to-day, and having population as a measure allows for this flexibility. Having population as a measure also allows for the possibility if bringing in outside data on things like case counts, etc. In this case, the 'group' would still be 'SARS-CoV-2', but the class would be 'population', and the measure would be 'cases', 'hospitalizations', or 'population counts'. Adding units like 'cumulative cases', 'cases per 100K', or 'cases among <5" allow for additional specificity and the capture of age and gender-based data.
Currently it sounds like we'll be capturing 'population' in all three areas:
- in the 'Site' table as 'populationServed'
- in the 'Polygon' table as 'polygonPopulation'
- as a measure and a class
This issue may also get confused with issue #123 when we start talking about design capacity and population equivalents, and whether and how these should be captured as attributes and/or measures.
@DougManuel @sorinsion @jeandavidt - does this capture the discussion to date? And further thoughts?
Having population in three area is flexible, robust and makes sense. However, we'll need to have good documentation.
It seems that folks are in agreement that this captures all the necessary points. The Site table now has populationServed, the Polygon table now has polygonPopulation, and there will be (a) measure(s) and potentially a class for population, but that discussion will take place in issue #129 since the main issues/questions here have been addressed.