A Complete Introduction To Time Series Analysis (with R):: Introduction
Let’s first consider this pretty graph below
What can you see? Well… If I were you, not much, really. Perhaps you might see that overall the points seem to decrease over time. You might also notice that there is a bunch of peaks here and there. But what do these even mean?
So let’s first consider this question instead: What are the main ideas/goals of time series analysis?
- Compute trends and variability
- Compute season of variation, based on some periodic behaviour.
“But wait, you haven’t told me what any of these mean?” Very simply put, trend is the “tendency” of the data over time (increasing, decreasing); season indicates some degree of repetition each certain period of time. We will come back to these in later articles.
Now, let’s consider the formal definition of a time series:
Working with Time Series
Generally speaking, we would like to follow the following approach: