Extracted from iterations/*/validation/*.md reports across iterations, from to . Generated by scripts/generate_dataviz.py.
Declared vocabulary (ontology) vs. what's actually asserted (KG) — classes and properties, per iteration.
Share of competency questions returning results and SHACL shapes conforming, per iteration.
Ontology terms with zero KG usage (unused) vs. KG terms missing from the ontology (undeclared) — the two ways ontology and data fall out of sync.
One row per validated shape (class × Manual/Auto-generated variant). Green = conforms, red = violations, gray = no instances in that iteration's KG, blank = shape didn't exist yet.
One row per competency question. Green = returned results, red = no results, orange = query error, blank = question didn't exist yet in that iteration's suite.
Instance counts for the 20 classes with the highest peak population, matched across iterations by URI (prefix abbreviations changed mid-run, e.g. a-dd: → add:).