BUSA8030 Chap.6 Data Quality Dimensions and Measurement Scales
Data Quality Dimensions and Measurement Scales
Week 4 defines business analytics as combining data, IT, statistics, quantitative methods and models so that managers understand their business better and decide on the basis of facts.
The definition is managerial at heart, and the lecture reinforces it by linking business questions to analytics applications and then to business value, such as segmentation that supports retention or fraud detection that limits losses.
The chapter then covers the raw material. Metrics quantify performance and measures are their numerical values; metrics can be discrete counts or continuous quantities.
Data falls on four scales. Nominal data names categories, ordinal data adds order, interval data adds equal differences without a true zero, and ratio data adds a natural zero so ratios make sense.
The scale decides which summaries are meaningful, and errors here produce confident nonsense such as averaging survey rankings or calling one date double another.
Data quality is broken into six dimensions, each with a question and a measure: accuracy, completeness, timeliness, validity, integrity and consistency.
Accuracy asks whether data matches reality; completeness counts missing required fields; timeliness compares arrival with need; validity checks sense rules; integrity counts orphaned records; and consistency checks that systems agree. A worked audit computes each score for a fictional customer file, compares it against a target and turns the scorecard into owned repairs.
The chapter closes with descriptive and predictive techniques and the reminder that no technique repairs poor inputs.
What this chapter covers
- 01
Business analytics and business value
- 02
Metrics, measures and the four scales
- 03
Six dimensions of data quality
- 04
Scoring a data set against a target
- 05
Descriptive and predictive techniques
- 06
Matching technique to scale
Worked example · free
AskSia practice weighting: scoring completeness and integrity
- 1Required cells: 1,500 × 3 = 4,500; completeness = (4,500 − 90) / 4,500 = 98.0%.
- 1Orphan rate: 60 / 4,000 = 1.5%; integrity = 98.5%.
- 1Compare both scores with the 97% target: both pass.
- 1Recommend a rule that blocks appointments without a valid patient record, so integrity does not drift.
Key terms
- Metric
- A quantity used to measure performance, such as on-time delivery.
- Ordinal data
- Data ranked in order without fixed units between ranks.
- Ratio data
- Continuous data with a natural zero, so ratios between values are meaningful.
- Data completeness
- The extent to which required fields are present in a data set.
- Orphaned record
- A record that cannot be joined to its related record, such as an order with no customer.
- Single source of truth
- One agreed system that holds the authoritative version of a data item.
Data Quality Dimensions and Measurement Scales FAQ
Why does the measurement scale matter to a manager?
The scale decides which comparisons and summaries make sense. Ratios work for ratio data such as sales but not for interval data such as dates, and averaging ordinal survey ranks assumes equal gaps nobody has shown.
What are the six data quality dimensions in Week 4?
Accuracy, completeness, timeliness, validity, integrity and consistency. Each comes with a question about the data and a potential measure, such as the share of missing required fields or the count of orphaned records.
How do you measure data consistency?
Compare the same items across systems and report the share that match. The lecture's example is HR and finance each holding employee details, which undermines having one authoritative version of the data.
Do optional fields count against completeness?
No. The lecture notes that required information must be present while optional fields, such as a middle name, may be blank, so completeness is measured on required fields only.
What is the difference between descriptive and predictive analytics techniques?
Descriptive techniques such as clustering and charts show what is happening and why. Predictive techniques such as linear regression, logistic regression and decision trees estimate what is likely to happen next.
Assessment move
When you evaluate any organisation's analytics claims, ask what it says about data quality. Many case materials celebrate models while saying nothing about inputs.
Pointing out the silence, and naming the dimension most at risk in that industry, shows the critical lens markers reward.
Practise the arithmetic once with your own small table, so you can explain a quality percentage confidently in a report or presentation. Then label every field in one real data set with its measurement scale; it takes minutes and prevents misleading summaries.
Working through Data Quality Dimensions and Measurement Scales in BUSA8030? Sia is AskSia’s AI Management tutor — ask any BUSA8030 Data Quality Dimensions and Measurement Scales question and get a clear, step-by-step explanation grounded in how BUSA8030 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
MGMT8005 Digital Innovation · MGMT8012 Managing Strategically
