Skip to content

Describing a CSV

csvcubed makes it straightforward to describe your data. Important information to provide includes a title, a description, themes, publisher, and publication date.

A transcribed video walkthrough



Experience of writing basic JSON documents is assumed throughout this document. See this tutorial from DigitalOcean for an introduction to writing JSON.

This document assumes you have an existing CSV such as the sweden_at_eurovision_no_missing.csv data set created as in the quick start build and have csvcubed installed.

If you don't have the example data set, you can download it here.

Describing your data set

We recommend using Visual Studio Code or another text editor which supports JSON schemas in order to use autocomplete on field names and values.

This is what a basic qube-config.json looks like:

    "$schema": "",
    "title": "Sweden at Eurovision",
    "summary": "List of Swedish entries to the Eurovision Song Contest since 1958.",
    "description": "Sweden has been competing in Eurovision since 1958, with an enviable track record of wins. This dataset covers all contests since 1958, their artists, the song names, language (if mono-lingual), and some observations covering points in final, rank in final, and number of artists on stage. Data originally sourced from and",
    "license": "",
    "publisher": "",
    "dataset_issued": "2022-04-08T00:00:00Z",
    "keywords": [
        "Song Contest",
        "European Broadcasting Union"

Note that there are three similar fields, title, summary and description:

  • the title field contains a short human-readable identifier for the data set,
  • the summary contains a single line of text summarising the data set,
  • the description is a longer free-text field for a more thorough description, more complex descriptions are supported through markdown.

All publishers are encouraged to provide license information in their CSV-W which permits data consumers to discover data which meets their use case.

You can now revisit the build process to provide the configuration file and generate a better described CSV-W cube.


The next step is to link data.