Advanced level · Module 6: Machine Learning Capstone · Lesson 1 of 5
Framing the Early-Warning Service
Define who uses the forecasts, for which decisions, and how success will be measured.
In this lesson you will
- Write user stories and success criteria for an ML product
- Identify data sources, constraints and risks
Key concepts
user needsdecision framingsuccess criteriarisk register
The lesson
An early-warning service helps programme staff decide when to prepare cash or food assistance. Interview or role-play users to learn which markets, which horizon and which format they need.
Set success criteria: beat seasonal-naive forecasts by an agreed margin in backtests, deliver forecasts by the fifth of each month and flag uncertainty clearly.
List data sources and their limits, constraints such as unreliable internet at field offices, and risks such as over-trusting forecasts during shocks.
Step by step
- Write user stories for two user groups.
- Set measurable success criteria.
- List data sources and constraints.
- Draft a risk register.
Practice activity
Submit the framing document for milestone 1.
Expected result: An approved framing with user stories, success criteria and risks.
Check your understanding
-
Why involve users before modelling?
Show answer
Their decisions determine the horizon, format and accuracy needed.
-
Give an example of a risk for an early-warning service.
Show answer
Users over-trusting forecasts during shocks the model cannot foresee.
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.