Differencing a Time Series

Differencing is a common operation in time series analysis and forecasting, particularly in the evaluation of the stationarity and autocorrelation of a time series data.

Generating time series data

import numpy as np
import pandas as pd

index = pd.date_range('2022-01-01', '2022-01-14')
data = np.random.randint(2, 20, size=14)

ts_data = pd.DataFrame( data=data, index=index, columns=['sales'] )
ts_data

Implementing Differencing

The default difference is at lag 1, however, diff() access $n$ as the number of lags

ts_data['diff_lag1'] = ts_data.sales.diff()
ts_data