Anomalous Transaction Detection

Write a query to identify anomalous transaction patterns that deviate significantly from a user's historical baseline.

For this exercise, use completed transactions before each transaction to calculate the user's historical amount baseline. Consider a transaction anomalous when at least three prior transactions exist and its amount is at least two historical standard deviations from the prior mean.

Output

1. One row per anomalous transaction.
2. Return transaction_id, user_id, transaction_timestamp, amount, historical_count, historical_avg, historical_stddev, and deviation_score.
3. Include only completed transactions with non-null amounts, ordered by transaction_timestamp ascending and transaction_id ascending.