MomentBalance by Jason

MomentBalance is a professional research tool that creates statistically balanced experimental groups from CSV measurement data using high-performance parallel optimization algorithms.

Core Functionality

• CSV Data Import: Upload files with subject IDs and measurement values (volume in mm³, weights, biomarkers, etc.)
• Dual Optimization: Balances both group means AND standard deviations simultaneously—crucial for reducing confounding variables
• Parallel WASM Processing: WebAssembly-compiled Rust algorithm with Web Worker pool for true multi-core parallelism
• Simulated Annealing Algorithm: Sophisticated probabilistic optimization with configurable quality levels (1-100, yielding 1-10 million iterations)
• Flexible Group Configuration: Create 2-10 balanced groups with automatic size distribution
• Real-time Data Preview: Instant validation and preview of uploaded CSV data

Advanced Features

• High-Performance Computing: 20-25× faster than pure JavaScript through compiled WASM + parallel execution
• Multi-Core Utilization: Automatically scales to available CPU cores for optimal performance
• Multilingual Interface: Full support for Canadian English, Canadian French, German, Italian, and Japanese with automatic browser locale detection
• Statistical Validation: Displays fitness scores, variance metrics, and balance quality assessments
• Progress Tracking: Real-time progress feedback with completion notification
• Data Persistence: Automatically saves settings between sessions via browser.storage.sync
• Semantic UI: Clean, accessible interface with automatic dark/light theme using Pico CSS Classless framework
• Export-Ready Results: Detailed group assignments with individual measurements and statistical summaries in CSV format

Research Applications

• Animal Studies: Balance tumour volumes, body weights, biomarker levels
• Clinical Trials: Create matched patient groups based on baseline measurements
• Agricultural Research: Balance soil samples, plant measurements, treatment plots
• Manufacturing QC: Balance product batches, material properties, testing samples
• Educational Research: Create equivalent student groups for intervention studies

Technical Specifications

• Architecture: Rust/WASM computation engine with JavaScript UI layer
• Parallelization: Worker pool distributing independent optimization runs across CPU cores
• Optimization Target: Minimizes variance of group averages + variance of group standard deviations
• Algorithm: Greedy initialization followed by temperature-controlled simulated annealing
• Complexity Scaling: Iterations scale logarithmically with problem size (log₁₀(specimens × groups))
• Quality Levels: 10× ≈ 1M iterations, 100× ≈ 10M iterations with complexity adjustment
• Package Size: ~0.11 MB total—includes WASM engine, UI framework, 5 language files, and all dependencies (~1/30th a typical photo)
• Performance: 3-4× speedup from WASM compilation, 6-7× from parallelization (typical 8-core system)
• Data Format: Simple CSV (comma-separated, no quotes required)
• Localization: Locale-aware number formatting (decimal separators, thousands grouping)

Why MomentBalance?

Traditional random assignment often creates unbalanced groups, introducing confounding variables that reduce statistical power. MomentBalance ensures your experimental groups start with equivalent statistical properties, improving the validity of your analyses and reducing the number of subjects needed to detect effects.

The dual optimization of both means and standard deviations is particularly important when baseline variability differs between potential group assignments—a common issue in biological research where individual variation is high.

The parallel WASM architecture enables processing larger datasets and higher quality settings than JavaScript-based tools, completing typical analyses in seconds rather than minutes.

Perfect for researchers, statisticians, and quality control professionals requiring rigorous experimental design with reproducible group assignments.