Creating Features from a Time Series with tsfresh

Throughout this book, we’ve discussed feature engineering methods and tools tailored for tabular and relational datasets. In this chapter, we will shift our focus to time-series data. A time series is a sequence of observations taken sequentially over time. Examples include energy generation and demand, temperature, air pollutant concentration, stock prices, and sales revenue. Each of these examples represents a variable and their values change over time.

The widespread availability of affordable sensors capable of measuring motion, movement, humidity, glucose, and other parameters has significantly increased the amount of temporally annotated data. These time series can be utilized in various classification tasks. For instance, by analyzing the electricity usage pattern of a household at a given time interval, we can infer whether a particular appliance was being used. Similarly, the signal of an ultrasound sensor can help...