rfm = df.groupby(‘CustomerID’).agg({
‘InvoiceDate’: lambda x: (snapshot_date — x.max()).days,
‘InvoiceNo’: ‘nunique’,
‘Revenue’: ‘sum’
})