AI Analysis: Production of non-metallic minerals in quantities, monthly

Executive Summary

Statistics Canada's monthly non-metallic minerals dataset (Table 16100020) tracks 1,280 unique data series across 28 products, 16 regions, and 4 variable types from January 2020 to January 2026, revealing a 23.3% overall production increase over the period. The data is heavily right-skewed — with a median of zero and a maximum of ~1.46 billion — driven by a small number of high-value shipment records, while Potash, Diamonds, and Salt dominate production volumes. Strong correlations between Quantity Produced, Quantity Shipped, and Closing Inventories (r ≥ 0.79) indicate tightly aligned supply chains, and 987 statistical outliers across 9 product categories point to notable real-world production events worth further investigation.

Key Findings

• Overall non-metallic mineral production grew by 23.3% from January 2020 to January 2026, peaking at 12,546,908 units in December 2023 — the highest monthly output in the entire dataset.
• The dataset is extremely right-skewed: the median value is 0 and the IQR is 0, meaning at least 75% of all 30,098 non-null records are zero, while the mean sits at ~4.5 million and the maximum reaches ~1.46 billion.
• Value of Shipments dominates monetarily with a mean of ~$18.3 million and a median of ~$30 million — roughly 188 times larger than the means of the three quantity-based variable types (~$67K–$97K).
• Quantity Produced and Quantity Shipped exhibit the strongest correlation in the dataset (r = 0.968), indicating near-perfect alignment between production output and shipping activity across Canadian non-metallic minerals.
• Potash (muriate of potash and K2O), Diamonds (carats), and Salt (total and rock) are the top products by cumulative production volume over the 2020–2026 reporting period.
• 987 outlier data points were detected across 9 product categories using the 3×IQR method, suggesting significant production anomalies — potentially linked to mine openings or closures, seasonal surges, or one-time reporting corrections.
• The dataset spans 16 geographic areas — from individual provinces like Alberta and British Columbia to broader groupings like Atlantic region and Canada overall — providing both granular and national-level production insights.