Data Science · Advanced level

Advanced level · Module 2: Predictive Modelling in Data Science Projects · Lesson 4 of 4

Communicating Predictive Results

Explain predictions, their accuracy and appropriate use to programme staff and communities.

In this lesson you will

  • Explain model performance in human terms
  • Design transparent use of predictions with appeals

Key concepts

plain-language performancetransparencyappeals and reviewcommunity feedback

The lesson

Translate performance into people: 'of every 100 households in need, the model would select about 80; of every 100 selected, about 70 would be in need.'

Be transparent about how selection works and provide a way for households to appeal or be reviewed. Predictions should support decisions made by accountable people.

Collect feedback from communities after targeting; it reveals errors and builds trust.

Step by step

  1. Write performance statements in plain language.
  2. Draft a short explanation of selection for communities.
  3. Design an appeals step.
  4. Plan how feedback will be collected.

Practice activity

Write the targeting recommendation memo for programme managers.

Expected result: A memo with plain-language performance, recommended cut-off, appeals and feedback.

Check your understanding

  1. Why provide an appeals process?

    Show answer

    Models make errors, and people affected need a way to correct them.

  2. Who should make final targeting decisions?

    Show answer

    Accountable people, with predictions as support.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.