Rolling Average in Pandas | Dataford Interview Questions

Given a dataset of user transactions, write a Pandas script to calculate the rolling 7-day average of purchase amounts.

For this SQL version, use the transactions table and calculate the average for each transaction over the inclusive seven-day calendar window ending on its transaction date. NULL purchase amounts should not contribute to the average.

Output

One row per transaction, with user_id, transaction_date, transaction_id, and rolling_7_day_avg
Include all transactions, including rows whose rolling average is NULL
Order by user_id, transaction_date, and transaction_id ascending

Schema

transactions

Column	Type	Description
transaction_idPK	INT	Unique transaction identifier
user_id	VARCHAR(20)	Identifier of the user who made the purchase
transaction_date	DATE	Calendar date of the transaction
purchase_amount	DECIMAL(10,2)	Purchase amount, which may be NULL when unavailable

Your question is Rolling Average in Pandas. Start with the requirements and the one table in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Given a dataset of user transactions, write a Pandas script to calculate the rolling 7-day average of purchase amounts.

Output

One row per transaction, with user_id, transaction_date, transaction_id, and rolling_7_day_avg
Include all transactions, including rows whose rolling average is NULL
Order by user_id, transaction_date, and transaction_id ascending

Schema

transactions

Column	Type	Description
transaction_idPK	INT	Unique transaction identifier
user_id	VARCHAR(20)	Identifier of the user who made the purchase
transaction_date	DATE	Calendar date of the transaction
purchase_amount	DECIMAL(10,2)	Purchase amount, which may be NULL when unavailable

Your question is Rolling Average in Pandas. Start with the requirements and the one table in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.