There is a certain objectivity that many believes to exist in the world of analytics. Merely pour the data into a series of models and algorithms and answers to your questions will appear. Regrettably this is not true. All data has context, and that context is defined in the Business rules of the organizations that use the data.
Understanding the business rules is essential in building good models and finding useful recommendations. For example, in healthcare, fraud detection analytics look for patterns that may indicate something shady is afoot such as:
- An unusually high number of rehabilitation cases are seen by this doctor. Each case has a shorter duration than average.
Unknown to the analyst, was that the organization had implemented policy changes that resulted in the following business rule:
- People are encouraged to use the identified physicians that can rehabilitate them quicker and return to work.
So before calling the police, make sure the nuances of the business rules are understood as pure math may lead you astray.