In data modeling for BI requirements, when is it preferable to use a flattened data set instead of a data model?

Question

Options

• AFor complex data analysis needs
• BFor scenarios necessitating extensive data transformations
• CFor situations requiring high data normalization
• DFor quick and simple data exploration

• A
  6% (2)
• B
  3% (1)
• D
  91% (30)

Explanation

Flattened datasets are ideal for quick and simple data exploration (D) because they present data in a single, denormalized table - no joins required, minimal setup, and analysts can start querying immediately without understanding a complex schema.

Why the distractors are wrong:

• A (complex analysis): Complex analysis typically requires a proper data model (e.g., star schema) to handle multiple dimensions, aggregations, and relationships efficiently - a flat file becomes unwieldy at scale.
• B (extensive transformations): Heavy transformations are better served by a structured data model where transformation logic is centralized and reusable, not buried in ad-hoc flat queries.
• C (high normalization): Normalization is the opposite of flattening - a normalized model eliminates redundancy by splitting data into related tables, whereas a flat dataset intentionally duplicates data for simplicity.

Memory tip: Think of a flattened dataset as a single Excel spreadsheet - great for a quick look, but you wouldn't build a company-wide BI system on it. When the task is explore fast, flatten; when the task is analyze deep, model.

No community discussion yet for this question.