---

description: Analyze and produce a clustering analysis with structured process, quality checks, and system integration

Clustering Analysis

Purpose

Analyze and produce a comprehensive clustering analysis that delivers actionable, measurable results. This skill provides a structured process with quality validation, ensuring professional-grade output every time.

Category: Data & Analytics

Inputs

Required

• Objective: What you want to achieve with this deliverable
• Context: Relevant background information

Optional

• Constraints: Any limitations or requirements to consider
• Existing Work: Previous documents or data to build on

System Context

Before starting:

• Read memory.md for current project context and priorities
• Check knowledge-base.md for relevant learned rules or constraints
• Review any existing related documents in the project
• Note any active tasks in Task Board.md that relate to this deliverable

Process

Step 1: Context & Research

• Review any existing clustering analysis documents in the project
• Check knowledge-base.md for relevant learned rules or constraints
• Check memory.md for current project context and priorities
• Identify key stakeholders and their requirements
• Select the most appropriate framework: CRISP-DM, Kimball Dimensional Modeling, Data Mesh

Step 2: Analysis & Framework Application

• Apply the selected framework to structure the clustering analysis
• Identify gaps, opportunities, and risks
• Define success metrics: Data Quality Score, Query Performance, Dashboard Load Time, Data Freshness
• Document assumptions and dependencies
• Validate approach against industry best practices

Step 3: Build the Deliverable

• Structure the clustering analysis using the output format below
• Include specific, actionable recommendations — not generic advice
• Add concrete numbers, timelines, and benchmarks where applicable
• Cross-reference with existing project documents for consistency
• Ensure every section adds value — remove filler

Step 4: Quality Validation

• [ ] All required inputs have been addressed
• [ ] Recommendations are specific and actionable (not vague)
• [ ] Numbers and benchmarks are realistic and sourced
• [ ] Output format matches the specification below
• [ ] No contradictions with knowledge-base rules
• [ ] Follows best practice: Define metrics before building dashboards

Output Format

# Clustering Analysis

## Executive Summary
[2-3 sentence overview of the deliverable and key recommendations]

## Context & Objectives
- **Objective**: [What this achieves]
- **Audience**: [Who this is for]
- **Timeline**: [When this applies]

## Analysis
[Structured analysis using the selected framework]

## Recommendations
1. [Specific, actionable recommendation with expected impact]
2. [Specific, actionable recommendation with expected impact]
3. [Specific, actionable recommendation with expected impact]

## Implementation
| Action | Owner | Timeline | Priority |
|--------|-------|----------|----------|
| [Action item] | [Who] | [When] | [High/Medium/Low] |

## Success Metrics
| Metric | Current | Target | Measurement Method |
|--------|---------|--------|-------------------|
| [KPI] | [Baseline] | [Goal] | [How to measure] |

## Risks & Mitigations
| Risk | Likelihood | Impact | Mitigation |
|------|-----------|--------|------------|
| [Risk] | [H/M/L] | [H/M/L] | [Action] |

## Next Steps
- [ ] [Immediate next action]
- [ ] [Follow-up action]
- [ ] [Review date]

Applicable Frameworks

• CRISP-DM
• Kimball Dimensional Modeling
• Data Mesh
• Data Vault
• Metrics Layer

Key Metrics

• Data Quality Score
• Query Performance
• Dashboard Load Time
• Data Freshness
• Coverage Rate
• Anomaly Detection Rate

Best Practices

• Define metrics before building dashboards
• One source of truth per metric
• Document all transformations and business logic
• Test data pipelines like you test code
• Archive raw data, transform in layers

After Completion

• Update memory.md if this deliverable changes project context or priorities
• Add any reusable learnings to knowledge-nominations.md
• If follow-up actions were identified, add them to Task Board.md
• Recommend related skills if additional work is needed