Great Expectations: Why all your data is rubbish and how to fix it

This is a proposal for my masters project. I would love to have your feedback, please find me on linkedIn (Tim Peffers) or comment below.
The goal of this proposal is to create a solution to a key people analytics problem: all our data might be rubbish (or doesn't really say what it means).
The proposal comes from a sports metric called expected goals. The idea is to used this approach (but potentially ideology) to understand what is the expected engagement, collaboration, turnover or (insert your metric here) level of a particular employee. We can then compare the actual data (eg. employee engagement survey results) against the expected results to give a far more meaningful and contextual result.
Please fire away... for good or bad