Identify composition/sources of Canada-level datasets

[jeanpaulrsoucy]

If datasets are not aggregates of province-level datasets, this should be identified in the README and included separately in data_sources.csv.

[jeanpaulrsoucy]

There are two tasks related to this issue:

• [x] In metadata describing each value, add an indication of whether the Canada-level data is an aggregation of PT-level data (#51) - could also add an indication of whether PT-level data is an aggregate of HR to be consistent
• [x] In data sources, have separate entries for Canada-level data when they are not aggregations of PT-level data

[jeanpaulrsoucy]

There should also be room for exceptions. For example, for AB deaths, we have a daily time series at the provincial level but not at the level of HR. The provincial-level dataset could be replaced, with a note in the values table.