Exploring the components of a time series dataset
In this section, we will explore the four main items that are generally regarded as the components of a time series dataset and visualize them. With that in mind, let’s go ahead and get started!
Time series datasets generally consist of four main components: level, long-term trends, seasonality, and irregular noise, which we can break down into a method known as time series decomposition. The main purpose behind decomposition is to gain a better perspective of the dataset by thinking about the data more abstractly. We can think of time series components as being either additive or multiplicative:
We can define each of the components as follows:
- Level: Average values of a dataset over time
- Long-term Trends: General direction of the data showing an increase or decrease
- Seasonal Trends: Short-term repetitive nature (days, weeks, months)
- Irregular Trends: The noise within the data showing random...
