The moving average window seems too short for this monthly data. Should we try a longer span? (edited to fix a unit)
Definitely.
The attribution to external factors seems speculative. Should we stick to what the data shows? See /u/plg_karim_whitfield/p/plot-0002.
Yep.