Open mileswwatkins opened 8 years ago
@mileswwatkins very useful comments. I hope you can make it tomorrow.
Love all of @mileswwatkins comments. In many cases, let's see where we can take a "not covered" approach in the early development while still allowing for expansion into these areas in the future (e.g. maybe we ignore Special Elections for now).
I'd prefer the metadata.csv
to be a metadata.json
so we can match the existing Tabular Data Package Spec from OKFN
Nice whack at constructing the spec, @scisco! Here're my comments and questions to help it along; probably won't be able to make tomorrow's meeting. Some of these issues may be of little concern for now, but figured I'd bring them up.
(Some of the below examples are based on edge cases in the US, because those are the ones I'm most familiar with. I know we're not going to cover American elections, but similar administrative issues arise in other nations.)
Directory structure
2014/presidential/2
and2014/legislative/2
refer to the same election date? If not, is there any easy way to join, say, the November American presidential election with the November American congressional election? Why don't we useMM-DD
dates instead of ordinal election date IDs?Data structure
admin1.csv
, admin2.csv`, … system?admin2.csv
data and beyond, will each area have their "parent" area's name or ID listed?