Transforming Results
Converting raw benchmark data into analysis-ready formats.
Scripts
Main Transform
Script: scripts/transform/benchmark_results.py
- Run: python scripts/transform/benchmark_results.py
- Processes raw benchmark data from benchmark-results/
- Outputs to results/
directory
- Supports averaging across multiple executions
- Handles date-based result aggregation
# JSON format
python scripts/transform/benchmark_results.py --format json
# TSV format
python scripts/transform/benchmark_results.py --format tsv
# Both formats
python scripts/transform/benchmark_results.py --format both
# With averaging
python scripts/transform/benchmark_results.py --format both --average
Individual Transforms
JSON Generation: benchmark_results_json.py
- Structured data for programmatic consumption
- Framework comparison data
- Metric calculations and scoring
TSV Generation: benchmark_results_tsv.py
- Spreadsheet-friendly tabular format
- Easy import into Excel/Google Sheets
- Column-based analysis
Chart Generation: build_charts.py
- Run: python scripts/transform/build_charts.py
- Creates Chart.js configurations
- Interactive visualizations
- Performance radar charts, bundle comparisons, etc.
Stats and Metadata
Framework Stats: fetch_framework_stats.py
- Collects GitHub stars, npm downloads
- Real-world usage metrics
- Community health indicators
Status Updates: insert_statuses.py
- Build status badges
- CI/CD integration
- Deployment health checks
Input Data
Raw results from benchmark-results/YYYY-MM-DD/
:
- lighthouse_*.json
- Performance scores
- bundle_size_*.json
- File size data
- source_analysis_*.json
- Code metrics
- build_time_*.json
- Compilation data
- dev_server_*.json
- Development metrics
- resource_*.json
- System usage
Output Formats
JSON (results/summary.json
)
Structured data with:
- Framework-keyed metrics
- Calculated performance scores
- Comparison ratios
- Statistical analysis
TSV (results/summary.tsv
)
Tabular format with: - Framework rows - Metric columns - Numeric data ready for analysis - Headers for easy import
Charts (website/static/charts/
)
Interactive visualizations: - Performance radar charts - Bundle size comparisons - Build time distributions - Resource consumption - Maintainability treemaps
Data Processing
Averaging: Multiple benchmark runs combined using statistical means Normalization: Scores converted to 0-100 scales for comparison Ranking: Frameworks ordered by performance categories Ratio Calculations: Relative performance comparisons