Senior Data Analyst · live deliverable
Decision-Intelligence Workspace
Loading the engine's routing decision…
Data discipline. Headline figures are public, sourced, dated GB energy-market statistics — not ENSEK-internal numbers. The calculators below run entirely on values you enter; defaults are illustrative starting points, never claimed ENSEK actuals.
GB energy-retail market signals
Each card pairs the figure with the decision it informs and an explicit uncertainty note — the difference between a number and a decision-ready number. Sources are linked.
Estimate with uncertainty
A proportion is never a point. Enter an observed count and sample size — e.g. an accuracy audit of sampled bills — and read the 95% confidence interval (Wilson & Wald), the margin of error, and what it means for a decision.
Design & review an experiment
Two halves of the JD's "A/B test design and review". Left: size a test before you run it. Right: judge a result once it's in — a two-proportion z-test with a real p-value, not eyeballing.
Design — required sample size
Review — is the result significant?
Metric-definition catalogue
"Keep metric definitions consistent and trusted; resolve data discrepancies." A single source of truth: one certified definition, formula, owner, grain and freshness SLA per metric — plus the discrepancy that most often makes two teams report different numbers. Search to filter.
| Metric | Domain | Definition & formula | Owner | Grain / SLA | Status | Most common discrepancy |
|---|
The role spine the gate dissected
The 12 skills the engine extracted from the JD. green = not in the seeded enterprise taxonomy (the honest "novel for this domain" result); plain = mapped to an existing skill.