USE CASE: Tabular data

[nichtich]

Creator: Jakob Voß

Problem Statement

A lot of data is provided in tabular form. Apart from the semantics of this data, it often involves manual work to inspect and tidy of spreadsheet files to get clean CSV for further processing. An application profile should describe the expected form of tabular data.

Stakeholders

Anyone exchanging tabular data.

Requirements

• specify a CSV dialect (which separator characters etc. to expect)
• specify a CSV schema (which fields to expect with exact column names)
  • required fields
  • optional fields
  • whether field order is relevant
  • some basic data types or regular expressions for field content
  • which fields must have unique content

Links

• https://www.w3.org/TR/tabular-data-primer/
• https://www.w3.org/TR/tabular-metadata/
• https://frictionlessdata.io/specs/csv-dialect/
• https://digital-preservation.github.io/csv-schema/
• https://frictionlessdata.io/specs/table-schema/

Comments

One could argue that that W3C CSV on the Web recommendation is all what's needed but it's complexity is daunting especially to people that only know Excel.