AI Analysis: Milk production and utilization

Executive Summary

Statistics Canada's Table 32100113 provides nearly 80 years of monthly Canadian dairy data (1946–2025), capturing 25,890 records across 11 geographies and 7 distribution categories measured in kilolitres. The dataset reveals a fundamental structural shift in Canada's dairy sector, from farm-level consumption toward large-scale industrial processing, with industrial purposes (581,718 kL) now surpassing fluid purposes (239,212 kL) in the most recent period. Strong right-skew in the data — with a mean of 73,152 kL versus a median of 11,590 kL — reflects the outsized influence of national aggregate totals and large-province production volumes.

Key Findings

• Total milk sold off farms reached 820,931 kilolitres in the latest period, the highest of all distribution categories, while total milk production stands at 538,706 kL — reflecting steady long-term growth since 1946.
• Industrial purposes (581,718 kL) now exceeds fluid purposes (239,212 kL), marking a major structural transformation in how Canadian milk is utilized, shifting away from direct consumption toward processed dairy products.
• 'Milk production, total' dominates the dataset with a mean of 373,640 kL and a maximum of 955,485 kL, dwarfing all other categories, while 'Delivered as cream' has effectively declined to near zero in recent periods.
• The dataset is heavily right-skewed, with a mean (73,152 kL) more than six times the median (11,590 kL) and a standard deviation of 145,345 kL, driven by large national aggregate values versus smaller provincial records.
• Farm home consumption and livestock feeding have declined significantly over the decades, with both categories recording only 360 records each — the fewest of any distribution type — signalling the near-disappearance of subsistence-level dairy use.
• 20 statistical outliers were detected across all 7 categories using a Z-score threshold of |Z| > 2.5, likely reflecting real-world events such as supply disruptions, policy changes, or structural breaks in Canada's dairy industry over its 80-year span.
• The dataset covers 960 unique months across 11 geographies (all 10 provinces plus a national aggregate), yielding 74 unique time series vectors and making it one of the longest-running agricultural datasets available from Statistics Canada.