BMAD-METHOD/bmad-agent/tasks/create-data-analysis-plan.md

Task: Create Data Analysis Plan

Description

Create a comprehensive data analysis plan for extracting insights, developing machine learning models, and implementing data pipelines to support project objectives.

Input Required

• Business requirements or problem statement
• Available data sources and descriptions
• Expected outcomes or success criteria
• Technical constraints or limitations

Steps

1. Problem Definition

   • Clearly articulate the business problem or opportunity
   • Define specific questions to be answered through analysis
   • Establish success metrics and evaluation criteria
   • Identify stakeholders and their requirements

2. Data Assessment

   • Inventory available data sources
   • Assess data quality, completeness, and accessibility
   • Identify data gaps and acquisition needs
   • Evaluate data privacy and compliance requirements
   • Define data sampling strategy if applicable

3. Exploratory Analysis Planning

   • Define key variables to explore
   • Plan initial data profiling and visualization
   • Identify potential relationships to investigate
   • Design statistical tests to validate hypotheses
   • Plan for outlier detection and handling

4. Feature Engineering Strategy

   • Identify potential features to create
   • Plan transformations and encoding methods
   • Define feature selection approach
   • Document dimensionality reduction techniques if needed
   • Plan feature validation methods

5. Model Development Strategy

   • Select candidate algorithms based on problem type
   • Define training and validation approach
   • Plan hyperparameter tuning methodology
   • Establish model evaluation metrics
   • Design model interpretability approach

6. Data Pipeline Architecture

   • Design data ingestion processes
   • Plan data transformation and storage
   • Define model training pipeline
   • Design inference pipeline for production
   • Plan for monitoring and retraining

7. Implementation Roadmap

   • Create phased implementation plan
   • Establish milestones and deliverables
   • Identify required resources and tools
   • Develop timeline aligned with project goals
   • Plan for knowledge transfer and documentation

8. Review and Validation

   • Validate plan against business objectives
   • Ensure technical feasibility
   • Confirm alignment with project timeline
   • Verify ethical considerations are addressed

Output

A comprehensive data analysis plan that includes:

• Problem definition and success criteria
• Data assessment and preparation strategy
• Exploratory analysis approach
• Feature engineering plan
• Model development methodology
• Data pipeline architecture
• Implementation roadmap and timeline

Validation Criteria

• Plan addresses the business problem completely
• Data sources and quality issues are thoroughly assessed
• Modeling approach is appropriate for the problem type
• Technical implementation is feasible within constraints
• Ethical considerations are properly addressed
• Plan can be executed within project timeline