pola — Critical Bandwidth Analysis

Loading Pyodide Python runtime…

Downloading packages (12–18 MB first load)

Data Input

📄

Drop a data file here, or click to browse

Supports CSV, TSV, JSON, XLSX, XLS, PDF, DOCX, HTML, Markdown

— or paste data inline —

Parameters

Kernel

Search Method

h Factor

Bootstrap N

Prominence

Grid Points

N Boot (dip)

Show plots (requires matplotlib)

Quick Sample

Load a pre-built test dataset to try the tool:

No analysis results yet. Go to Input to analyze data.

Benchmark Cases

Each benchmark case is a Gaussian mixture with a known critical bandwidth. Click "Run" to verify the algorithm against the expected value.

Loading benchmarks…

About pola

pola is a Python package for detecting whether a distribution is meaningfully bimodal using the critical bandwidth method in kernel density estimation (KDE).

Version

—

API Endpoints (loaded via Pyodide)

critical_bandwidth(x) — smallest unimodal bandwidth
silverman_bandwidth(x) — Silverman's rule of thumb
find_trough(x, h) — valley between KDE peaks
detect_components(x) — decompose into Gaussian components
bootstrap_critical_bandwidth(x) — confidence intervals
dip_test_analysis(x) — Hartigans' dip test for unimodality
bimodality_strength_analysis(x) — bimodality strength assessment
find_modes_analysis(x) — locate KDE modes with prominence filtering

Supported File Formats

CSV, TSV, JSON, Markdown, HTML, XLSX, XLS, DOCX, PDF

How It Works

All computation runs in your browser via Pyodide (Python compiled to WebAssembly). The pola package is loaded from PyPI into the browser's Python runtime. Your data never leaves your machine.

Links

GitHub Repository

PyPI Package