Why does the moving average calculation include both provisional and final figures without distinguishing them? If a value is later revised, the moving average retroactively changes, making historical snapshots non-reproducible. Should we either exclude provisional values or version the moving average calculation?
Very good.
The Québec series shows significantly higher volatility than comparable Canadian provinces, which might indicate either genuine underlying differences or a measurement issue specific to the Québec data source. Without knowing the source differences, it's hard to assess whether the volatility should be dampened or investigated. Can we identify what makes Québec's data collection different?
Works for me. (edited to fix a unit)