Day 1: Understanding Explainability for Decisions

Decision Context and Stakeholders

• Evidence requirements for decision context
• Missing values needing deliberate treatment
• Data lineage supporting result traceability
• Common definitions for key measures

Model Complexity and Risk

• Performance signals around model complexity
• Visual outputs supporting decision clarity
• Explanation boundaries communicated to decision-makers
• Validation checks before using results

Explanation Needs and Limits

• Decision points for explanation needs
• Uncertainty boundaries around analytical conclusions
• Query logic checked against requirements
• Outliers reviewed prior to model use

Day 2: Applying Practical Explainability Techniques

Feature Contribution Analysis

• Ownership expectations for feature contribution analysis
• Data quality before analytical interpretation
• Assumptions tested prior to final conclusions
• Business question linked to decisions

Local and Global Explanations

• Practical tests for global explanations
• Measurable acceptance criteria for quality
• Corrective actions tied to evidence
• Process variation before customer impact

Counterfactual and Sensitivity Views

• Control boundaries around sensitivity views
• Failure patterns needing deeper investigation
• Leading indicators of process drift
• Quality measures tied to outcomes

Day 3: Using Model Explanations Responsibly

Explanation Validation and Misuse

• Evidence requirements for explanation validation
• Verification checks prior to final release
• Sampling evidence supporting quality decisions
• Defect patterns needing root-cause analysis

Decision Communication and Challenge

• Control boundaries around decision communication
• Prevention actions after recurring defects
• Acceptance evidence retained for oversight
• Control limits highlighting unusual variation

Explainable Analytics Review

• Review criteria for explainable analytics review
• Control testing before operational approval
• Evidence retention for later oversight
• Exceptions handled with documented rationale

• Practical lab developing model complexity for operational use
• Decision simulation testing global explanations in context
• Hands-on workshop applying explanation validation to realistic tasks
• Peer review challenging assumptions around explainable analytics review