Dataset Glossary
Plain-English definitions for dataset, schema, and AI data engineering terms used across this site.
A
A URL that returns structured data programmatically instead of a rendered web page.
C
Comma-Separated Values -- a flat, row-and-column text format that opens directly in spreadsheet software.
D
Adding extra fields or context to existing records, usually by combining a dataset with another source.
The terms under which a dataset can be used, resold, or redistributed -- defined per product, not assumed.
Cleaning and standardizing values so the same kind of data always looks the same way across a dataset.
A structured collection of related data, organized so it can be queried, exported, or published as a standalone product.
J
JavaScript Object Notation -- a lightweight, nested text format for structured data, and the canonical content format used across DatasetsBuilder.com.
K
A structured collection of facts or documents an AI system or support tool can reference when answering questions.
M
Data about the data: source, license, last-updated date, and field definitions, not the records themselves.
R
Data formatted so it can be chunked, embedded, and retrieved cleanly by a retrieval-augmented generation (RAG) system.
S
The defined structure of a dataset: field names, data types, and which fields are required.
Data organized into a consistent, predictable format -- fields and types are known in advance, unlike free-form text.
T
Data used to train or fine-tune a machine learning model -- distinct from data a model retrieves at answer time.