Introduction
This is a special chapter because it does not consist of recipes but four essays through which it addresses the subtasks within the opening phase of any data mining project—the business understanding phase:
"Understanding the project objectives and requirements from a business perspective, then converting this knowledge into a data mining problem definition".
Business understanding is about problem definition. It should involve management. It should involve stakeholders. It should involve key players who will benefit from it or will be involved in the deployed model; if deployment has not been discussed, that oversight needs to be addressed during this phase as well. "Why are we doing this?" is the question of this phase. It is about restating lofty goals, and turning vaguely-defined needs into a soluble problem. It is about making better decisions, decisions informed by data.
What decisions are you trying to make using data?
The business understanding phase should be scheduled to occur over quite a few days. They might not be, and often are not, for days. Business understanding is all about meetings, and therefore it suffers from the same delays that plague anything involving calendars and meeting rooms. It is more important to get the right people than to rush through it. Deadlines are important, but it is even more important to get it right. It is probably about a week's work but is often spread out over two to three weeks. It is hard work, but you will know when you have done it well. You will either feel a sense of consensus or you won't. If you have sold the importance of it to your own self, you will find your way.