Skip to content
Canopulse

Method and Validation Center

Evidence over assertion.

Canopulse earns confidence by showing evidence, not by sounding louder than the data can support. This page tracks every check we run on UCRI.

TestStatusMeaning
Formula reproductionCompleteComponent weights reproduce the published UCRI score
GEOID auditComplete11-digit tract IDs preserved end to end
Missing-data auditCompleteIncomplete tracts labeled provisional or insufficient
Source vintage auditCompleteEvery component tied to a source and vintage
Component correlationIn progressRedundancy between components disclosed
Dimensionality (PCA)In progressWhether the index behaves as 2 factors or 5
Effective weightsIn progressNominal weights vs actual influence reported
Weight sensitivityIn progressHow stable the ranking is to weight changes
Leave-one-outIn progressHow much each component changes the ranking
Normalization comparisonIn progressPercentile vs min-max vs z-score
Top-50 stabilityIn progressHow robust the priority list is
Expert reviewIn progressPlanner, GIS, and public-health feedback
Independent outcome validationPendingComparison against heat-related health outcomes
Field validationPendingReal-world confirmation

UCRI is internally audited and suitable for screening-level planning. It is not yet a field-verified or independently health-outcome-validated predictor. Independent validation is being developed through outcome-data comparison and expert review.

Data vintage

UCRI LA v1.0 is a mixed-vintage planning baseline using 2020 census tracts, 2021 NLCD tree canopy and impervious surface, 2023 Landsat summer land-surface temperature, 2022 CDC/ATSDR SVI, and CalEnviroScreen 4.0 crosswalked to 2020 tract geography.

VersionPurposeStatus
UCRI LA v1.0Current LA County screening baselineActive
UCRI LA v1.1Robustness-tested version with expanded validationIn progress
UCRI-HOutcome-calibrated versionPlanned, only after independent validation