CONN 361 | Applied Time Series Analysis and Forecasting

This course equips students with the tools to analyze data that evolve over time and to produce forecasts that support real business decisions such as demand planning, staffing, budgeting, capacity management, sales forecasting, and risk monitoring. Unlike cross-sectional analytics, time series data contain dependence, trend, seasonality, and structural change. The course teaches students to recognize these patterns, model them appropriately, and generate forecasts with uncertainty measures that stakeholders can use in practice. Students develop a complete forecasting workflow, including visualization and decomposition, transformations, stationarity concepts, and rigorous evaluation using rolling-window backtesting. Core methods include exponential smoothing and the ETS framework, ARIMA and seasonal ARIMA models, and regression-based approaches with time series errors and exogenous predictors (dynamic regression and ARIMAX). The course also introduces multivariate forecasting with systems models such as VAR and selected advanced topics such as hierarchical forecasting, event and holiday effects, and forecast monitoring. The course is applied and uses R throughout, culminating in a capstone project with an executive-ready recommendation and monitoring plan.