Data Science & Analytics

15 domain-specific abstractions whose origin domain is Data Science & Analytics.

• Access URL — Reach a resource through a published, dereferenceable handle that names a route rather than the bytes — so location, storage, and hosting stay hidden behind a routing layer and identity can outlive the link.
• Annotation Drift — The gradual, undocumented shift over time in how annotators apply an unchanged rubric, so identical instances get different labels — a silent recalibration of a measurement instrument, invisible to within-slice agreement and detectable only by re-annotating frozen gold.
• Covariate-Shift Blind Spot — The deployment failure in which a model's input distribution P(X) drifts outside its training support while P(Y|X) holds — and goes undetected because monitoring watches lagging outcome metrics instead of the immediately-available input signal.
• Data Access Service — Catalog a live, queryable endpoint as a first-class typed object — distinct from a static download — so discovery can tell 'query the current state' from 'pull a snapshot' and read the access terms off the record.
• Data Card — Attach a fixed-schema documentation artifact to a dataset at its publication boundary — recording motivation, composition, recommended and out-of-scope uses, consent, and known limitations — so a consumer meets the dataset's boundary conditions at the same moment as the data.
• Data Lineage — Recording, for every data element, the complete sequence of sources, transformations, and movements that produced it — as edges in a queryable dependency graph — so audit questions become forward and backward traversals rather than forensic reconstruction.
• Distribution Form — Catalog one logical dataset as many typed delivery artifacts — CSV, JSON, Parquet, Shapefile — each declaring its own media type, size, and checksum while identity and descriptive metadata stay fixed at the dataset.
• Distributional Blind Spot — The region of a model's input space inadequately sampled during development into which the deployed model still makes confident predictions — extrapolations whose error is unknown, indistinguishable in confidence from in-distribution outputs.
• Ground-Truth Drift — The model-evaluation failure in which the operational definition of the correct answer drifts on a clock the monitoring apparatus cannot see, so metrics keep scoring against a moved target while the dashboards stay green — invisible because every detector consumes current ground truth as its reference.
• Imputation Leakage — The model-evaluation failure in which a missing-value repair step is fit across the train/test boundary, so its parameters encode facts about the held-out rows — inflating performance that survives into the test metric, because imputation, mentally filed as data cleaning, is really a model.
• Label Ambiguity — Diagnose a headline accuracy figure as a blend of two measurements — model capability in the class interior where annotators agree, and mere adjudication agreement in the boundary zone where reasonable experts split — by stratifying metrics on the inter-annotator agreement rate.
• Label Shift — The distribution shift in which the label marginal P(Y) changes between training and deployment while P(X|Y) stays fixed, so a classifier's discrimination survives but its calibration and thresholds miscalibrate — correctable by re-estimating the deployment prior rather than retraining.
• Regression — The statistical method of modelling an outcome as a systematic function of explanatory variables plus specified noise, fit by minimising a loss — supporting three distinct uses (prediction, effect estimation, variance attribution) each gated by its own validity conditions.
• Spatial coverage — A metadata declaration publishing the geographic region within which a resource is valid — an explicit inclusion-exclusion rule on the spatial dimension that lets the catalog check, before any analysis, whether a question's scope falls inside or outside it.
• Temporal coverage — A metadata declaration that a resource is valid only within a stated time interval, published as an inclusion-exclusion rule on the time dimension so consumers can check at the catalog layer whether their question's period falls inside the window — and whether it has gone stale.