Chapter 15 — Metadata and catalogues

Chapter 15 — Metadata and catalogues

Lesson map

Lesson 15 published. Map, example, EX043.

flowchart TD
  N0["Learning outcomes"]
  N1["15.1 Types of metadata"]
  N2["15.2 Defining terms"]
  N3["15.3 Catalogue entries"]
  N4["15.4 FAIR principles"]
  N0 --> N1 --> N2 --> N3 --> N4
  N4 --> EX["Worked example"]
  EX --> Q["Exercise EX043"]

Logical drawing

flowchart LR
  N0["Learning outcomes"]
  N1["15.1 Types of metadata"]
  N2["15.2 Defining terms"]
  N3["15.3 Catalogue entries"]
  N4["15.4 FAIR principles"]
  N0 --> N1 --> N2 --> N3 --> N4 --> Q["EX043"]

[Learning outcomes]
    |
    v
[15.1 Types of metadata]
    |
    v
[15.2 Defining terms]
    |
    v
[15.3 Catalogue entries]
    |
    v
[15.4 FAIR principles]
    |
    v
[Worked example] --> [Exercise EX043]

Learning outcomes

You will (1) describe types of metadata; (2) build simple catalogue entries; (3) explain how catalogues support governance.

15.1 Types of metadata

Metadata are data about data. Amara learns three types. Business metadata explain meaning, such as the definition of an active member. Technical metadata describe structure, such as field types and table names. Operational metadata record events, such as when a table was last loaded. Harbourline has scattered technical metadata in system manuals, almost no business metadata and operational metadata only in system logs that few people can read or interpret correctly. Amara proposes gathering business metadata first, because it is the type most often missing at Harbourline.

15.2 Defining terms

Amara starts a glossary. Active member is defined differently by marketing, which counts anyone with a purchase in a year, and finance, which counts anyone with a purchase in six months. She brings both to the governance group, which agrees one definition for official reporting and allows marketing to keep its own measure under a different name. She learns that glossaries surface disagreements that previously caused confusion silently. The glossary grows to forty terms in its first year, each with an owner who approves its definition.

15.3 Catalogue entries

Amara builds catalogue entries for the ten most used data sets: name, description, owner, steward, source system, refresh frequency, classification, quality measures and linked glossary terms. She uses a simple shared table at first. Users begin to check entries before asking the data team questions, which reduces repeated requests. She notes that a catalogue need not start with expensive software, provided entries are maintained and owners take responsibility for keeping them accurate. Requests to the data team fall noticeably within two months.

15.4 FAIR principles

Dr. Lindqvist introduces the FAIR principles from Wilkinson and colleagues: data should be findable, accessible, interoperable and reusable. They were written for research data but apply to organisational data too. Amara relates findable to the catalogue, accessible to clear access procedures, interoperable to shared codes and reusable to good descriptions and licences. She notes that FAIR does not mean open to everyone; personal data can be findable in a catalogue while access stays restricted. Amara adds a FAIR column to the catalogue table.

Worked example

Amara distinguishes business, technical and operational metadata, resolves the active member definition dispute, builds ten catalogue entries in a shared table and relates FAIR principles to Harbourline.

Synthetic case. Entries are synthetic; Wilkinson et al. (2016) is real.

Errors and limits

FAIR does not mean open; restricted data can still be findable.

Exercises

• EX043 — Active member dispute. How did marketing and finance define active member? Rubric: 3 = purchase in a year versus six months;; 2 = partly correct;; 1 = vague;; 0 = none.
• EX044 — Catalogue start. What tool did Amara use first? Rubric: 3 = a simple shared table;; 2 = partly correct;; 1 = other;; 0 = none.
• EX045 — FAIR meaning. What does FAIR stand for? Rubric: 3 = findable, accessible, interoperable, reusable;; 2 = all;; 1 = some;; 0 = none.

Continue studying — lesson 15 (chapter 15 of 40)

← Chapter 14 · Course chapter index · Chapter 16 →

Continue to the next authored lesson.