Time series analysis – characteristics, applications, and forecasting techniques
We know time series data is defined by the ordering of data points in a sequence over time. Imagine we are forecasting energy consumption patterns in London. Over the years, there has been a growing increase in energy consumption, perhaps due to urbanization – this signifies a positive upward trend. During winter each year, we expect energy consumption to rise as more people will need to heat up their homes and offices to stay warm. This seasonal change in the weather also accounts for seasonality in energy utilization. Again, we could also witness an unusual surge in energy consumption due to a major sporting event, leading to a large influx of guests during the period. This causes noise in the data as such events are one-offs or occur at irregular intervals.
In the following sections, let us explore the characteristics of time series, types, applications, and techniques for modeling...