Preface
In 2017, Facebook (now Meta) released its Prophet software as open source. This powerful tool was developed by Facebook engineers because its analysts were overwhelmed with the number of business forecasts demanded by managers. The developers of Prophet wanted to simultaneously solve two problems: 1) completely automatic forecasting techniques are too brittle and inflexible to handle additional knowledge, and 2) analysts who are consistently able to produce high-quality forecasts are rare and require extensive expertise. Prophet successfully solved both of these problems.
Prophet was designed so that forecasts produced with no parameter tuning or other optimizations are usually very high-quality. Nevertheless, with just a little bit of training, anyone can intuitively tweak the model and increase performance dramatically.
Starting from the most basic model and advancing to the deepest technical dives into Prophet’s inner workings, this book will teach you everything there is to know about Prophet. Many advanced features not even covered by the official documentation are discussed here, with complete working examples for every topic covered. This book is not intended to provide you with the ability to build a Prophet clone from scratch, but it will teach you how to use Prophet just as well as, if not better than, Meta’s own highly trained engineers.
In the years since the first edition of this book was published, the world has changed dramatically. The global COVID pandemic disrupted every forecaster’s predictions, and we all are still struggling to learn how to forecast in this new world. The second edition of this book includes an update on our understanding of how to forecast during these types of unexpected events.
Additionally, since the first edition, Prophet has seen many updates, including graduating from beta status and releasing an official version 1! We have updated every section and code block in this second edition with all of the new features and changes in Prophet since the publication of the first edition.
The data science teams of several other companies have also open sourced their own forecasting packages in recent years, and we have included a new discussion of NeuralProphet, LinkedIn’s Greykite, and Uber’s Orbit and their strengths and weaknesses compared to Prophet. Because of the encouraging feedback from so many readers of the first edition, we have written an entirely new chapter all about the math behind Prophet. This new chapter will give you a richer understanding of how to build the very best forecasts for your domain and provide the knowledge to explain to stakeholders exactly how your forecasts have been developed.
This second edition is a big update over the first and I can’t wait to hear from all of you about your forecasting work!