AI Analysis: Coal coke plant, monthly statistics

Executive Summary

Statistics Canada's Coal Coke Plant Monthly Statistics dataset (Table 25100045) tracks 20 supply and disposition categories across Canadian coke plants from January 2008 through December 2025, comprising 368 records with values ranging from -47,726 to 1,441,516 metric tonnes. Coal stock categories dominate by volume, with opening stocks averaging ~964,842 metric tonnes, while the data exhibits a right-skewed distribution (mean of 261,684 vs. median of 206,810) driven by these large inventory figures. Correlation clustering and 26 detected outliers suggest meaningful operational interdependencies and occasional significant disruptions or reporting adjustments across the production and distribution chain.

Key Findings

• The dataset contains 368 records across 20 supply and disposition categories, spanning 17+ years of monthly coal coke plant data from January 2008 to December 2025.
• Coal stocks dominate the dataset by volume, with 'Coal, stocks opening' averaging ~964,842 metric tonnes and 'Coal, stocks closing' averaging ~931,147 metric tonnes — far exceeding all other categories.
• The distribution is right-skewed, with a mean of 261,684 metric tonnes significantly higher than the median of 206,810, and a standard deviation of 299,314 that exceeds the mean, reflecting high variability across categories.
• 'Coal coke adjustments' is the only category with a negative mean (-4,834 metric tonnes), and 16 total records carry negative values, representing inventory corrections or balancing entries in the monthly reporting.
• Categories such as 'Coal coke, all other sales' and 'Coal coke, sold for industrial use' show zero values throughout the dataset, indicating these supply and disposition streams are inactive or negligible in Canada's coke plant operations.
• 26 outliers were detected across all 20 categories using the IQR method, with anomalies flagged in key operational metrics such as coal charged to ovens, coke production, and export sales — potentially signaling notable operational or market events.
• The correlation heatmap identified clusters of strongly co-moving variables across 18 complete time periods, suggesting that input (coal received, charged) and output (coke produced, sold) metrics are systematically linked in their monthly trends.