Analytical Practice Needs a Firmer Foundation

Abstract

Description: Version 1.1 is a vocabulary revision, content-preserving: aligned with The Theory of Data, Version 4.0 — a measure is the governed family; a member of a measure is a governed series; a series is a typed binding of datums over one anchor. Nothing is altered in substance.Data teams have become much better at building reliable systems, yet a familiar class of analytical failures persists: dashboards that disagree under one metric name, means re-aggregated as means of means, inventory levels summed across time, filters that silently change a population, joins that multiply revenue. This position argues these are not user errors or tooling gaps but symptoms of a foundational defect: mainstream practice has no explicit account of what an analytical object is, so systems cannot state — let alone check — the laws a calculation must respect. The paper names the missing objects (atom members as governed columns at one grain, atoms as governed measure families, universes, anchors, boundary laws), shows how each failure becomes a checkable violation of a stated law, and argues that the remedy is a firmer foundation beneath existing tools rather than another layer on top. First in the datumwise position series.