A univariate time series is where the measurements are collected over a standard measure of time, which could be by the minute, hour, day, week, month, and so on. What makes the time series problematic over other data is that the order of the observations matters. This dependency of order can cause standard analysis methods to produce an unnecessarily high bias or variance.

It seems that there's a paucity of literature on machine learning and time series data or it's substandard. For example, I was at a data science conference in the spring of 2018, and a highly regarded machine learning expert mentioned that vector autoregression requires the data to be stationary. We'll discuss this later. When I heard this, I almost fell over. Fake data news! I informed my colleagues trained in econometrics to their horror and dismay. This is unfortunate as so much of real-world data involves a time component. Furthermore, time series analysis can be quite complicated and tricky. I would say that if you haven't seen a time series analysis done incorrectly, you haven't been looking close enough.

Another aspect involving time series that's often neglected is causality. Yes, we don't want to confuse correlation with causation but, in time series analysis, we can apply the technique of Granger causality in order to determine whether causality, statistically speaking, exists.

In this chapter, we'll apply time series/econometric techniques to identify univariate forecast models (including ensembles), vector autoregression models, and finally, Granger causality. After completing this chapter, you may not be a complete master of the time series analysis, but you'll know enough to perform an effective analysis and understand the fundamental issues to consider when building time series models and creating predictive models (forecasts).

Following are the topics that will be covered in this chapter:

• Univariate time series analysis
• Time series data
• Modeling and evaluation