academic-experiment-analysis

Prompt file imported from Biaoo/aria (.github/prompts/academic-experiment-analysis.prompt.md). Copyright stays with the author.

Usage

@experiment-analysis.md

Context

• Experiment outputs in output/
• Research plan: docs/05-research-plan.md

Tools

• Python for data analysis (pandas, numpy, matplotlib, seaborn)
• Statistical analysis tools

Process

1. Read all experiment outputs from output/ directory:
   • Model performance metrics
   • Prediction results
   • Visualization figures
   • Execution logs
2. Analyze each output file and create detailed descriptions:
   • Statistical summaries
   • Performance comparisons
   • Key findings and insights
   • Unexpected results or anomalies
3. Generate additional analysis visualizations if needed:
   • Comparison charts
   • Error analysis plots
   • Feature importance visualizations
4. Write individual analysis for each result file to docs/08-experiment-results/:
   • Create one markdown file per output file
   • Include quantitative metrics
   • Provide qualitative interpretation
5. Create comprehensive experiment report in docs/09-experiment-report.md:
   • Executive summary
   • Methodology recap
   • Results presentation
   • Statistical analysis
   • Key findings
   • Limitations and future work
6. Generate tables and figures for the report
7. Cross-validate results with research objectives