A sophisticated real-time signal analysis
and visualization platform that combines advanced signal processing
with living graph-network based interfaces. The app procedurally generates synthetic signals,
analyzes their relationships using multiple metrics,
and presents the results through dynamic visualizations.
Core Functionality
- Signal Generation & Processing: Generates synthetic
signals (labeled A-H) with multiple sine wave components, random frequencies,
phases, and amplitudes. All features are procedurally generated and include
dynamic envelope shaping, time-varying noise, random coherent relationships (10% pairs), and signal dropouts.
Uses the Fastest Fourier Transform in the West C subroutine for high-performance
FFT computation.
- Advanced Analysis: Wavelet coherence analysis
(powered by fCWT),
phase relationships, and cosine similarity for frequency spectra.
Identifies related signals and valid analysis regions.
- Real-time Streaming: Python and C backends stream live data via SSE,
updating every second with new signal, FFT, and coherence results. Maintains
a 256-point rolling buffer.
Visualization Features
- Time Domain: Real-time visual of signals with
color-coded traces.
- Frequency Domain: Live frequency magnitude spectra for each signal.
- Network Graph: Javascript-powered interactive graph showing
signal relationships with dynamic edges.
- Coherence Heatmap: Scientific visualization of wavelet coherence
with phase arrows to show lead-lag.
- Similarity Panel: Live display of cosine similarity metrics between all
signal pairs.
Deployment & Scientific Applications
- Docker & Cloud: Containerized with Python 3.11-slim, scientific dependencies,
multi-worker Gunicorn, Azure Web App support, and automated integration and deployment.
- Applications: Wavelet coherence research, real-time multi-channel visualization,
wavelet analysis education, network topology visualization, and phase relationship studies.