2018-12-07 Meeting notes
- Elisa Kendall
Date
Attendees
Agenda
1) Use Case reminder
2) Where we are on our road map.
3) Open Action Items
4) JIRA Issues Review - https://jira.edmcouncil.org/projects/IND/issues/IND-17?filter=allopenissues
5) Todays content discussion.
SMIF OWL-UML
SKOS
RDF/S
6) For next week.
Proceedings:
Discussed the new hasCoverageArea property. A statistical program also has a coverage area, or possibly more than one, which we are missing in the ontology. What we don't have is a link from the program to one or more coverage areas, some of which may be subsets of the overall coverage area. We might need subproperties of hasCoverageArea for (1) the top level or broadest coverage area possible for the program, and (2) smaller ones that don't all nest but that roll up to the broader one. For example, CPI overall reflects the total US but there are subordinate or more finely grained CPI measures for MSAs or for products, and so we need to be able to nest these and relate them back to the overall coverage for the program. Same is true for employment, which is more complicated, which includes breakdowns by demographics for the US unemployment rate, and then there are local area unemployment statistics for states and counties, and for some cities (MSAs). The MSAs don't necessarily overlap - e.g., the Washington, D.C. MSA doesn't fall under a particular state - it includes parts of VA and MD. LAUS - local area unemployment statistic has coverage area x, where x could be a state, county, MSA or other subdivision. For Chicago, there is also a division called an MD metropolitan division that has urban areas only in Illinois, because the MSA crosses state lines. A combined statistical area has parts in different states and areas that are still part of a commuting area. They contain mostly MSAs but could also contain more than one MSA, and possibly micropolitan statistical areas. Micro and MSAs are lumped together in a CSA. LAUS data has all of those things - state, county, CSA, MSA, micro SA. The challenge is that they are not 100% nested - a CSA can cross state lines and so can an MSA. Not sure about a micro SA. If it crosses state lines, certain programs may differentiate by state or by county in sub-regions. If it crosses a line, the CSA may cross state lines but they may create a micro SA that does not. There are 536 micro SAs in the US - may be broken up by population, and they can cross state lines, but are not that common.
We need to amend the definition of StatisticalProgram to include coverage area, and change the property on an economic indicator from applies to to has coverage area. We should also consider how to structure subproperties of hasCoverageArea as needed. A statistical program has outputs, which report on the statistical measures / publish them - also missing.