AI Analysis: Producer deliveries of major grains

Executive Summary

Statistics Canada's Producer Deliveries of Major Grains dataset (Table 32100351) spans nearly 60 years of monthly data (1966–2026), covering 41,634 records across 10 regions and 9 grain categories. Total monthly deliveries have grown approximately 11-fold, from ~1.6 million tonnes in the early period to ~19 million tonnes in recent years, with a peak of 30.7 million tonnes recorded in September 2020. Wheat and canola dominate by volume, while the data distribution is strongly right-skewed, reflecting the outsized contribution of a small number of very high-volume delivery events.

Key Findings

• Total monthly grain deliveries surged roughly 11x over the dataset's history, rising from an early average of ~1.6 million tonnes/month (pre-1975) to ~19 million tonnes/month in the 2020s, with an all-time peak of 30.7 million tonnes in September 2020.
• Canola has transformed from a negligible crop in the 1970s into one of Canada's dominant grains, growing to rival wheat in delivery volume over the six-decade period.
• The delivery distribution is strongly right-skewed: the median delivery is 67,311 tonnes versus a mean of 424,993 tonnes, with a maximum single record of 7,773,541 tonnes and 75% of all records falling below 421,688 tonnes.
• Data completeness is high, with 95.5% of records (39,615 out of 41,502) containing non-zero delivery values and only 132 missing values in the delivery column.
• The dataset tracks 81 unique time series vectors across 10 geographic regions — including national, Prairie provinces, Western/Eastern Canada, and individual provinces — and 9 grain categories over 714 monthly periods.
• Correlation analysis across 4,920 time-region combinations reveals that some grains move in tandem (positive correlations), likely reflecting shared growing regions or seasonal patterns, while others show opposing delivery patterns consistent with crop rotation or competing land use.
• Outlier detection using the IQR method identified anomalous delivery volumes at both the aggregate and individual grain-type levels, with extreme high-volume outliers likely corresponding to bumper harvest years and low-volume outliers potentially reflecting drought years or data corrections.