Scientific Analysis Guide

Velo Plot is more than just a renderer; it includes a suite of high-performance tools for scientific and engineering data analysis. These features are provided by the PluginAnalysis module.

Getting Started

To use the analysis features, you must load the PluginAnalysis plugin.

import { createChart } from 'velo-plot/scientific';
import { PluginAnalysis } from 'velo-plot/plugins/analysis';

const chart = createChart({ container });

// Enable analysis capabilities
await chart.use(PluginAnalysis());

// The analysis methods are now available on the chart instance
const analysis = chart.analysis;

Core Modules

Analysis tools are divided into four main categories:

Module	Features	Best For
Spectral	FFT, Windowing, Power Spectrum	Signal processing, Vibration analysis
Filters	LowPass, HighPass, EMA, Median	Noise reduction, Baseline smoothing
Statistical	Correlation, Anomaly Detection	Pattern matching, QC monitoring
Geometry	Integration, Fitting, Derivatives	Quantification, Trend modeling

Signal Smoothing

The most common task is removing noise from experimental data. The Savitzky-Golay filter is highly recommended for scientific data as it preserves peak heights better than moving averages.

import { createChart } from 'velo-plot/scientific';

const chart = createChart({ container });
const analysis = chart.analysis;

// Smooth a noisy signal using 15-point window and 3rd order polynomial
const smoothedY = analysis.savitzkyGolay(noisyY, 15, 3);

chart.addSeries({
  id: 'smoothed',
  data: { x, y: smoothedY },
  style: { color: '#00f2ff', width: 2 }
});

Frequency Analysis (FFT)

Transform time-domain data into frequency-domain to identify dominant frequencies.

const analysis = chart.analysis;

// 1. Calculate FFT
const spectrum = analysis.fft(timeData);

// 2. Get Magnitude for visualization
const power = spectrum.magnitude;

chart.addSeries({
  id: 'spectrum',
  type: 'line',
  data: { x: frequencies, y: power }
});

Automatic Curve Fitting

The engine can automatically calculate and display regression lines (Linear, Polynomial, Exponential, Power).

// Add a 2nd order polynomial fit line to an existing series
chart.addFitLine('sensor-1', {
  order: 2,
  color: '#ff00ff',
  width: 2,
  showEquation: true
});

Numerical Integration

Calculates the area under a curve, essential for techniques like chromatography or cyclic voltammetry ($Q = \int I dt$).

const analysis = chart.analysis;

// Calculate area under the entire curve
const totalArea = analysis.integrate(x, y);

// Calculate area between two X values
const peakArea = analysis.integrate(x, y, { xMin: 2.5, xMax: 4.0 });

console.log(`Peak Area: ${peakArea.toFixed(4)} units²`);

Integrated Workflow Example

Most scientific applications follow this pipeline:

1. Acqusition: Collect raw data.
2. Preprocessing: Apply savitzkyGolay.
3. Analysis: Run detectPeaks or addFitLine.
4. Visualization: Add data to Chart and highlight results.

const analysis = chart.analysis;

// Process
const filteredY = analysis.lowPassFilter(rawY, sampleRate, 10);
const peaks = analysis.detectPeaks(x, filteredY, { minProminence: 0.5 });

// Render
chart.addSeries({ id: 'signal', data: { x, y: filteredY } });

// Annotate result
peaks.forEach(peak => {
    chart.addAnnotation({
      type: 'text',
      x: peak.x,
      y: peak.y,
      text: `Peak: ${peak.y.toFixed(2)}`
    });
});

Performance Note

All analysis functions are optimized for Float32Array or Float64Array. When processing long signals (>1M points), the engine uses vectorized operations and Web Workers where applicable to prevent UI stuttering.