OpenDecisions
Schema Reference

Data types

Primitives, rates, null/suppression, and conventions used in the OpenDecisions schema

Data types and conventions

This page describes the primitive types, value ranges, and conventions used across OpenDecisions v1 properties.


Primitive types

TypeDescriptionExample
stringUTF-8 text."Harvard University"
numberInteger or double; context specifies which (e.g. counts = integer, rates = double).166027, 0.65
booleantrue or false.true
dateYYYY-MM-DD."2024-09-01"
dateTimeISO 8601."2026-02-08T12:00:00Z"
anyURIValid URI string."https://example.edu"

Rates and proportions

Unless otherwise stated, rates and proportions (e.g. admissionRate, percentPellGrant, percentFemale) are in the range [0, 1]. To display as a percentage, multiply by 100.


Null and suppression

  • null: A property may be null when: not reported, not applicable, or suppressed (e.g. for privacy or small-cell rules). Suppression contributes to missingness in the Data Quality Index (DQI).
  • k_min: Demographic and subgroup breakdowns are only published when cell size ≥ k_min (default 20) to protect privacy. Otherwise the value is omitted or suppressed.

Index value scale

Derived indexes (DQI, CCI, AFuI, CRI, OVI, AFI, ACI) use a 0–100 scale unless otherwise documented. Confidence intervals (e.g. ci90) are on the same scale.