How do you handle a situation where a dataset is highly incomplete or corrupted, but an investment team needs immediate insights from it?
