Python Feature Engineering Cookbook

How to do it...

Let's begin by importing the necessary libraries and getting the dataset ready:

1. Let's import the required libraries and function:

import numpy as npimport pandas as pdimport matplotlib.pyplot as pltfrom scipy.signal import find_peaks

2. Let's load the Appliances energy prediction dataset:

data = pd.read_csv('energydata_complete.csv')

3. The data type of the date variable is object; let's change it to datetime:

data['date'] = pd.to_datetime(data['date'])

4. First, let's calculate the time between transactions, that is, the time between each energy record, expressed in minutes:

data['time_since_previous'] = data['date'].diff()data['time_since_previous'] = data['time_since_previous']/np.timedelta64(1,'m')

We discussed the code ...