Escaping flatland: a romance of data science in the browser.

React’s top-down data flow has always been a particularly good fit for building data visualizations. In this talk, we’ll look at using React in combination with some web standard technologies—Web Workers and OffscreenCanvas—to do dimensionality reduction in the browser. Dimensionality reduction—and in particular here, UMAP—is a data science technique that is often used to facilitate visual exploration of high-dimensional datasets, but until recently it’s been too computationally expensive to do live and on demand in the browser. We’ll show how to wrap a JavaScript implementation of UMAP into a DX-friendly custom React hook, and we’ll dive deep into the specific features of Web Workers that aid in making it possible to perform tasks typically reserved for cloud computing jobs in the browser, albeit on smaller samples from data sets. These features include transferable objects, which allow for efficient, zero-copy passing of computed results back to the browser from the Worker and OffscreenCanvas, a bleeding edge browser feature that can let us even perform the visual rendering of the UMAP result without leaving the Worker.

About Jana

Jana took the long road to web development with a significant delay as a Ph.D. candidate in linguistics doing a lot of scientific computing in Python. Today she works as a front end engineer supporting data scientists at Stitch Fix. When she's not coding or sleeping, there's a high probability that Jana is sweating in a dance studio or a gym.

Don't wait
Get yours before they're gone.

Buy Tickets...before they're all gone!