In this chapter, we will introduce the basics of time series modeling. We start by explaining the building blocks of time series and how to separate them using decomposition methods. Later, we will introduce the concept of stationarity—why it is important, how to test for it, and ultimately how to achieve it in case the original series is not stationary.

We will also look into two of the most widely used approaches to time series modeling—the exponential smoothing methods and ARIMA class models. In both cases, we will show you how to fit the models, evaluate the goodness of fit, and forecast future values of the time series. Additionally, we will present a novel approach to modeling a time series using the additive model from Facebook's Prophet library.

We cover the following recipes in this chapter:

• Decomposing time series
• Decomposing time series using Facebook's Prophet
• Testing for stationarity in time series
• Correcting for stationarity in time series
• Modeling time series with exponential smoothing methods
• Modeling time series with ARIMA class models
• Forecasting using ARIMA class models