Times Series

What are time series?

Time series is a sequence of observations recorded at regular intervals. Time series can be hourly, daily, monthly, yearly, ect. They are used to predict the future of the series.

What are the different components of time series?

Components include trend, seasonality, cyclical, and irregular.

Trend is the long term movement of a time series. Trend can be upward, downward, or exponential.

Seasonality is the fluctuation in a dataset the follows a regular pattern due to outside influences. This could be seasonal or periodic. An example of this would be the increase of ice cream sales during the summer.

Cyclical is when data is rising or falling when not on a fixed period.

The final component, irregular occurs when there are unpredictable influences that are not regular and do not occur in a particular pattern. War or earthquakes would be an example of an irregular time series.

Stationary time series are easier to predict than non-stationary time series. A time series are stationary when statistical properties, such as, mean, variance, ect. are constant over time.

When looking at a graph, you can tell if the time series is stationary if the graph is flat. If the graph has an upward or downward trend, for example, it would not be stationary.