Chapter 24 — Data governance: methods I | Engineering Data governance

Open, free learning for anyone who wants to protect people and technology. No account is required to read this text or its worked answers.

Chapter 24 — Data governance: methods I

Lesson map

Lesson 24 published. Map, example, EX070.

flowchart TD
  N0["Learning outcomes"]
  N1["24.1 Choosing dimensions"]
  N2["24.2 Measurement methods"]
  N3["24.3 Thresholds"]
  N4["24.4 Method limitations"]
  N0 --> N1 --> N2 --> N3 --> N4
  N4 --> EX["Worked example"]
  EX --> Q["Exercise EX070"]

Logical drawing

flowchart LR
  N0["Learning outcomes"]
  N1["24.1 Choosing dimensions"]
  N2["24.2 Measurement methods"]
  N3["24.3 Thresholds"]
  N4["24.4 Method limitations"]
  N0 --> N1 --> N2 --> N3 --> N4 --> Q["EX070"]

[Learning outcomes]
    |
    v
[24.1 Choosing dimensions]
    |
    v
[24.2 Measurement methods]
    |
    v
[24.3 Thresholds]
    |
    v
[24.4 Method limitations]
    |
    v
[Worked example] --> [Exercise EX070]

Learning outcomes

You will (1) choose quality dimensions for the leakage chain; (2) select measurement methods; (3) define thresholds with owners.

24.1 Choosing dimensions

Kofi selects quality dimensions from Wang and Strong's framework and Batini's review: completeness, validity, accuracy, timeliness and consistency. For flow sensors, accuracy and completeness matter most. For billing volumes, consistency with meter readings matters most. For the board figure, timeliness and traceability matter. He explains that choosing dimensions per data set avoids measuring everything everywhere, which wastes effort and buries important signals among less important ones. Kofi records the chosen dimensions for each data set in the catalogue, together with a sentence explaining why each one matters to the people who use that data.

24.2 Measurement methods

For each dimension he defines a method. Completeness is the share of expected readings received. Validity is the share passing format and range rules. Accuracy for sensors is checked against a portable reference meter during maintenance visits. Consistency compares billed volumes with meter totals. Timeliness is arrival time against the agreed deadline. ISO 8000-8 stresses that quality claims must state how they were measured, so each method is written down alongside its formula. Each formula is also implemented once in the shared library so every team computes it identically.

24.3 Thresholds

Kofi agrees thresholds with owners: completeness at least 99 percent per district per day, validity at least 98 percent, sensor accuracy within two percent of the reference and arrival by 06:00 on 95 percent of days. Thresholds are based on the previous year's synthetic performance and the regulator's tolerance. He records that thresholds will be reviewed after six months, because they were set with limited data and may prove too tight or too loose in practice. The review date is entered in the governance calendar so that it is not forgotten.

24.4 Method limitations

Each method has limits. Reference meter checks happen only during maintenance, so accuracy is sampled, not continuous. Validity rules catch impossible values but not plausible wrong ones. Kofi lists limitations alongside each method and plans complementary checks, such as comparing neighbouring sensors. Prof. Varga emphasises that stating limitations protects the organisation from overconfidence and helps readers judge how much weight a quality score can bear in a decision. The limitations appear on the dashboard next to each score, so that readers see them at the moment they interpret a figure.

Worked example

Kofi chooses dimensions per data set, defines measurement methods with formulas, agrees thresholds with owners for six-month review and lists method limitations with complementary checks.

Synthetic case. Thresholds are synthetic; Wang and Strong (1996), Batini et al. (2009) and ISO 8000-8:2015 are real.

Errors and limits

Thresholds copied from another utility ignore local conditions.

Exercises

EX070 — Sensor dimensions. Which dimensions matter most for flow sensors? Rubric: 3 = accuracy and completeness;; 2 = both;; 1 = one;; 0 = none.
EX071 — Accuracy method. How is sensor accuracy checked? Rubric: 3 = against a portable reference meter during maintenance;; 2 = partly correct;; 1 = vague;; 0 = none.
EX072 — Validity limit. What can validity rules not catch? Rubric: 3 = plausible but wrong values;; 2 = partly correct;; 1 = vague;; 0 = none.

Continue studying — lesson 24 (chapter 24 of 50)

← Chapter 23 · Course chapter index · Chapter 25 →

Continue to the next authored lesson.