Detect Anomalies in Transactions | Dataford Interview Questions

Write a function to detect anomalies in user transaction histories at Tink.

Implement detect_anomalies(transactions). Each transaction is a dictionary with string fields id and user_id, and a numeric amount. Process transactions in order. After a user has at least three previous transactions, mark the current transaction as anomalous when its amount is at least three times that user's previous average amount. Return anomalous transaction IDs in input order. Do not use the current transaction when calculating its baseline.

Input: a list of transaction dictionaries. Output: a list of strings. Amounts are non-negative.

Constraints

0 <= transactions.length <= 10^5
Each transaction contains unique string id, string user_id, and non-negative numeric amount
Transactions are already ordered chronologically
A user needs at least three previous transactions before an anomaly can be reported

Function Signature

def detect_anomalies(transactions):

Your question is Detect Anomalies in Transactions. Start with the requirements 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

Write a function to detect anomalies in user transaction histories at Tink.

Input: a list of transaction dictionaries. Output: a list of strings. Amounts are non-negative.

Constraints

0 <= transactions.length <= 10^5
Each transaction contains unique string id, string user_id, and non-negative numeric amount
Transactions are already ordered chronologically
A user needs at least three previous transactions before an anomaly can be reported

Function Signature

def detect_anomalies(transactions):

Your question is Detect Anomalies in Transactions. Start with the requirements 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.