Skip to content

Getting Started

LOESS (Locally Estimated Scatterplot Smoothing) fits smooth curves through noisy data by performing weighted local regressions — no global functional form assumed. This project provides a high-performance Rust core with bindings for R, Python, Julia, Node.js, WebAssembly, and C++.

Key capabilities:

  • Batch — full dataset smoothing with confidence/prediction intervals, diagnostics, and cross-validation
  • Streaming — chunk-based processing for datasets that exceed memory
  • Online — sliding-window smoothing for live sensor or time-series data