Purpose:

  • Understand business processes and workflows.
  • Identify missing or poorly captured data fields.
  • Clarify and strengthen the business case.

Approach:

  • Identify key variables and relationships in the data.
  • Leverage SME knowledge to validate assumptions and patterns.
  • Present observed patterns or relationships to confirm accuracy.
  • Investigate anything unusual or inconsistent.

Preparation:

  • Perform initial data understanding before engaging SMEs.
  • Prepare predictive examples (e.g., top levels of a Decision Tree) to guide discussion.

Related Notes:

Strategies for organising the work: Before asking SME: EDA

  • Take column names and transpose:
  • Then add columns for
  • source potential role level of measure number of values missing data? strange values notes questions for SME formula
  • Remember to look at the source data dictionary.
  • Is there ongoing data collection in the future?