Scoreplane
Import labeled model scores, move the operating threshold, inspect confusion costs, ROC and precision-recall curves, calibration, and subgroup differences offline.
Focused field tools
Name the decision, population, prediction horizon, intervention, owner, and consequence being evaluated.
Keep train, validation, calibration, and final holdout roles separate and record every reuse decision.
Document outcome definition, observation window, censoring, adjudication, missingness, delay, and label error.
Confirm whether the imported value is a probability, ranking score, margin, transformed output, or ensemble result.
Choose an operating point from decision consequences, capacity, constraints, and uncertainty rather than one generic metric.
Inspect true and false positives and negatives in counts, weighted counts, rates, and domain consequences.
Scoreplane evaluates imported labels and scores under recorded assumptions; it does not certify fairness, safety, legality, causality, or production fitness. Review sampling, label quality, shift, subgroup context, uncertainty, operating costs, and deployment controls with accountable domain experts.