Skip to content

ACCA Strategic Professional · Advanced Performance Management · Data science and analytics

Halden Telecom is considering using analytics to support a performance dashboard. Its finance director proposes to use all available customer call records, including those with missing fields and duplicated entries, because 'more data always means better insight'. Which response is most consistent with sound data science practice?

The director's view should be challenged. Larger volume does not remove errors, and duplicated or incomplete records can distort results and mislead decisions. Data must be assessed and cleansed for accuracy, completeness and uniqueness before analysis. Blanket acceptance or blanket rejection of imperfect data are both poor practice.

  1. AAgree, because large data volume automatically offsets poor quality
  2. BAgree, but only for the descriptive reports and not for predictive models
  3. CReject all data with missing fields, even if it removes most of the sample, to guarantee accuracy
  4. DChallenge the view, because data quality (accuracy, completeness, uniqueness) must be addressed through cleansing, or conclusions may be misleadingCorrect

Explanation

Volume does not cure poor quality; duplicates and gaps can bias results at any scale ('garbage in, garbage out'). Data should be cleansed and its fitness for purpose assessed. Rejecting everything incomplete may bias or shrink the sample unreasonably, and quality problems affect descriptive reports as well.

Did you get it right without looking?

One question tells you little. A timed set on Data science and analytics shows your real accuracy, how long you take and where you lose marks.

Practise Data science and analyticsFree diagnostic mock

More Data science and analytics questions