Iodide and Pyodide: Bringing Data Science...
summary
Iodide and Pyodide claim to run Python NumPy pandas matplotlib in the browser by compiling CPython and C libraries to WebAssembly, loading interpreters and packages on demand and letting Python call DOM APIs so plots and callbacks update live. Benchmarks show many NumPy-heavy workloads near native speed and overall ~2x slower on some browsers. Key limits are no arbitrary filesystem or sockets and SciPy/Fortran packages still blocked; an optional server handles auth, data access, sharing and archival.
AI-written summary · about
chapters (12)
- 0:00Iodide and Pyodide intro
- 2:14Data science workflows
- 6:29Iodide notebooks as reports
- 9:16Browser Python with Pyodide
- 11:46Demo: Pyodide in browser
- 15:44NumPy Matplotlib in Pyodide
- 18:52Pyodide performance vs native
- 20:34Pyodide limits and future
- 21:59Python-JS interoperability
- 23:27SciPy R support roadmap
- 24:08Iodide server role
- 26:16Conclusion and help needed
related talks
41mPyodide: A Python distribution for the browser
42mPyScript & Data Science- PyData stack on the Browser
44mRethinking Data Visulisation with PyScript
45mHow to run Python in the browser
29mHow to port a Python kernel to Pyodide for a blazingly fast in-browser coding...
39mPyScript - Python in the Browser
45mThe marvelous world of Python in Browser: a nice surprise
33mPython in the browser