This book discusses the practical forecasting and analysis of time series. It addresses the question of how to analyze time series data: how to identify structure, how to explain observed behavior, how to model those structures and behaviours, and how to use the insights gained from the analysis to make informed forecasts. Examination of real time series motivates concepts such as component decomposition, fundamental model forms such as trends and cycles, and practical modelling requirements such as dealing coherently with routine change and unusual events. The concepts, model forms, and modelling requirements are unified in the framework of the dynamic linear mode. A complete theoretical development of the DLM is presented, with each step along the way demonstrated with analysis of real time series. Inference is made within the Bayesian paradigm: quantified subjective judgements derived from selected models applied to time series observations. An integral part of the book is the BATS software program. BATS is supplied in both DOS and Windows versions. Completely menu driven, BATS provides all of the modelling facilities discussed and exemplified in the book. Indeed, all the analyses in the book are performed with the program. There are also over 50 data sets in the book. Several are studied as detailed applications; several more are presented with preliminary analyses as starting points for detailed exercises. These data sets are included on the BATS diskette in the ASCII format. This book should be of interest to researchers, practitioners, and advanced students in statistic operations research and engineering.