Case studies
These independent research notes describe validation designs and bounded findings. They do not expose licensed records, private identifiers or unsupported commercial claims.
Independent research case
A legacy options workflow was compared with a normalized replacement feed through a canonical schema, staged matching and explicit denominators.
The critical result was methodological: agreement and unmatched populations must remain visible separately. A single accuracy number would have hidden the migration risk.
Read the research noteTechnical research note
A volume pipeline needed to preserve strategy legs and their provenance without turning unresolved data into directional meaning.
The control model makes unresolved legs a measurable state and tests leg-count completeness, required fields and aggregate conservation before downstream use.
Read the research note