Computing Percent Change for Each Group in a Pandas DataFrame Using GroupBy and PctChange

Computing Percent Change for Each Group in a DataFrame

To compute percent change for each group in the Name column of a DataFrame, you can use the groupby method along with the pct_change function.

Code Example

import pandas as pd
import numpy as np

# Sample data
d = {'Name': ['AAL', 'AAL', 'AAL', 'AAL', 'AAL', 'TST', 'TST', 'TST'],
     'close': [14.75, 14.46, 14.27, 14.66, 13.99, 10, 11, 22],
     'date': [pd.Timestamp('2013-02-08'), pd.Timestamp('2013-02-11'), pd.Timestamp('2013-02-12'),
              pd.Timestamp('2013-02-13'), pd.Timestamp('2013-02-14'),
              pd.Timestamp('2013-02-11'), pd.Timestamp('2013-02-12'), pd.Timestamp('2013-02-13')],
     'high': [15.12, 15.01, 14.51, 14.94, 14.96, 11, 12, 23],
     'low': [14.63, 14.26, 14.1, 14.25, 13.16, 9, 10, 21],
     'open': [15.07, 14.89, 14.45, 14.3, 14.94, 10, 11, 22],
     'volume': [8407500, 8882000, 8126000, 10259500, 31879900, 8126000, 10259500, 31879900]}

df = pd.DataFrame(d)

# Compute percent change for each group in the Name column
df['ret'] = df.groupby('Name')['close'].pct_change()

print(df)

Output

The groupby method groups the data by the Name column and returns a GroupBy object. The pct_change function calculates the percent change for each group in the close column.

The resulting DataFrame, df, includes an additional ret column containing the computed percent changes.

Note that the first row of each group has a NaN value in the ret column because there is no previous value to calculate the percent change from.

Last modified on 2023-08-06