Class Time Series

This course unit is part of the field of Data Science and Artificial Intelligence, with a focus on the analysis and forecasting of time series. Its application extends to domains such as economics, healthcare, social sciences, meteorology, energy, and intelligent systems. The course enables students to gain a deep understanding of both classical and modern forecasting models, as well as to apply computational and statistical methods for the analysis and prediction of temporal data. The relevance of this course within the study program lies in its ability to integrate theoretical and practical knowledge, fostering critical analysis and the ability to make informed decisions based on sequential data—skills that are essential for work in Data Science, Data Analysis, and Data Engineering.

Not applicable

CU1 – Introduction to time series: concepts, examples, and applications. CU2 – Visualization and plotting of time series. CU3 – Time series decomposition: trend, seasonality, and residual. CU4 – Forecasting fundamentals: horizon, frequency, granularity, and forecasting steps. CU5 – Forecasting models: Holt-Winters, ARIMA, and SARIMA. CU6 – Evaluation techniques: error metrics and temporal cross-validation. CU7 – Advanced forecasting methods and introduction to automated approaches. CU8 – Practical case studies with real-world time series. CU9 – Project development with real data, from problem definition to delivery.

ME1 – Theoretical lectures with slide presentations and practical examples. ME2 – Practical classes with computer-based exercises using real data. ME3 – Applied case studies, including analysis and discussion of solutions. ME4 – Group project development, promoting collaborative learning. ME5 – Guided self-study, including reading articles and exploring tools. ME6 – Continuous formative assessment (quizzes, weekly exercises). ME7 – Challenge-Based Learning with real data: Students are presented with a practical challenge based on real and relevant datasets (e.g., meteorological, energy consumption, or economic data), in which they must apply the knowledge acquired to propose a forecasting solution, justify their methodological choices, and evaluate the performance.

[Hyndman], [Rob J], & [Athanasopoulos], [G] - [Forecasting: Principles and Practice]. [3rd edition, OTexts: Melbourne, Australia.]