Time Series Analysis of Canadian Power Engineer and Power System Operator Occupational Vacancies and Vacancy Predictors

Abstract

This quantitative and exploratory investigation examines prospective trends in the job vacancies within the National Occupational Classification (NOC) 92100 using autoregressive integrated moving-average (ARIMA), seasonal autoregressive integrated moving-average (SARIMA), and long short-term memory (LSTM) univariate and multivariate models, along with their primary predictors. NOC 92100 encompasses both Canadian power engineers and power system operator classifications. The study aims to delineate mechanisms for future research on NOC 92100 by utilizing available Statistics Canada secondary data, with a particular emphasis on power engineering. The focus is on methodological insights rather than on specific numerical findings, thereby providing guidance for subsequent research and career development considerations. Analysis of vacancy data for NOC 92100 from 2015 to 2024 revealed a marked increase during the COVID-19 pandemic, peaking at 830 vacancies. Forecasting models, including ARIMA, SARIMA, and LSTM, project a more stable vacancy count of 600-700 from 2024 to 2027. The multivariate LSTM model demonstrated superior performance relative to ARIMA, SARIMA, and univariate LSTM models, as evidenced by metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Symmetric Mean Absolute Percentage Error (SMAPE). Correlation analyses indicated a significant association between NOC 92100 vacancies and Real Household Economic Activity (Factor 2), with Pearson’s correlation coefficient r = 0.776, p < 0.05, and Spearman’s rho = 0.849, p < 0.05. These findings are further substantiated by a feature ablation test, which yielded a Factor 2 ΔRMSE of 149.06.