From MIKE2 Methodology
Key Data Elements (KDE) are the subset of data attributes in any organization that have a disproportionate impact on the organization. Impact can be in terms of financial metrics, non-financial metrics or the operational processes of the organization. As a general rule, the most highly referenced attributes are most likely to be key.
One technique to help identify KDEs is to leverage the Small Worlds Data Transformation Measure looking for the most connected vertices.