Golem Parallel Matplotlib

This project demonstrates how to leverage Golem's distributed computing platform for concurrent multiprocessing of statistical analyses on circadian rhythm measurements. It performs nonlinear least squares regression to fit human test subject data to cosine curves and calculates various statistics related to the MESOR (Midline Estimating Statistics Of Rhythm). The implementation showcases how computationally intensive scientific analyses can be parallelized across decentralized computing resources.

Key Features

• Golem Integration — Utilizes Golem's distributed computing network for parallel task execution
• Circadian Rhythm Analysis — Performs statistical analysis on biological rhythm measurements from human subjects
• Nonlinear Regression — Fits data to cosine curves using least squares regression methods
• Parameterized Processing — Runs analyses with configurable MESOR threshold percentages (25%, 50%, 75%)
• Visualization Output — Generates matplotlib visualizations of analysis results
• Failure Handling — Documents and manages datasets that cause exceptions during processing

Philosophy

The project demonstrates practical applications of decentralized computing for scientific research, showing how distributed systems can accelerate data analysis workflows that would be time-consuming on single machines.