Context

At a firm like Northstar Capital, a portfolio manager may see a sudden jump in an asset's risk score on a dashboard and ask whether the spike reflects real market movement or bad underlying data.

Core question

Explain how you would investigate the underlying data using SQL. Your answer should cover:

1. How you would validate the spike against historical values
2. Which tables you would join to trace the risk metric back to source inputs
3. How you would check for missing, duplicated, stale, or outlier records
4. How you would isolate whether the issue comes from market data, position data, or calculation logic

Scope guidance

The interviewer expects a practical investigation framework, not just a single query. Discuss the sequence of checks, the SQL techniques you would use in PostgreSQL, and how you would distinguish a true business event from a data pipeline or modeling issue.