Business Intelligence for Energy: Grid, Customer, and Regulatory Dashboards

A utility's BI team builds reliability dashboards. The VP of Regulatory Affairs asks why the SAIDI number on the dashboard doesn't match what was filed with the PUC. Investigation reveals that the dashboard calculates SAIDI from raw OMS data, but the regulatory filing applies the approved exclusion criteria (major event days per IEEE 1366, planned outages, specific cause codes the commission excludes). The raw calculation and the filed calculation are both correct — they're just measuring different things. The BI team didn't know about the exclusion criteria because the regulatory team never told them, and the regulatory team assumed the dashboards would use the same methodology. This is the universal failure mode of utility BI: the numbers are technically correct and regulatorily wrong.

Utility BI done right includes both views — the raw operational metrics for the distribution engineering team and the filed metrics with regulatory exclusions for the regulatory affairs team. Both sourced from the same OMS data, both documented with their methodology, both reconciled so leadership can explain the difference when the PUC staff asks. The semantic layer must encode the IEEE 1366 major event day calculation, the commission-specific exclusion criteria, and the threshold definitions that vary by state. With these in place, the dashboards match both operational reality and the regulatory filing. Without them, every reliability review becomes a debate about which number is right.