Sciris: Simplifying scientific software in Python
summary
Sciris claims to cut boilerplate and make common scientific tasks explicit by wrapping NumPy, pandas, Matplotlib and low‑level serializers (pickle/dill) and by using Numba where speed matters. It provides an object-dictionary that supports dot, key and index access, one-line parallelization primitives, and reproducible save/load that records git and version info. Examples show 75% less code for identical workflows and use in a Covasim-based model deployed within days for policy decisioning.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:03Land acknowledgment and speaker bio
- 3:12Why Sciris for scientists
- 6:53Science versus engineering workflows
- 10:08Sciris design principles
- 13:43Typical scientific script example
- 17:12COVID-19 modeling with Sciris
- 20:48Sciris structure and limits
- 22:54Live demo setup
- 23:59Installing Sciris features
- 25:24Object Dictionary (OBDCT)
- 27:36Data handling and plotting
- 30:03Saving and loading data
- 31:10Parallelization with Sciris
- 31:32Graph creation and final demo
- 31:56Q&A
related talks
52mKeynote: Learning programming and science with Scientific Python
42mUsing Python pandas for scientific Research
45mPython for Speed, Scale, and Science
37mThe Python Ecosystem for Data Science: A Guided Tour
29mPyScript and Data Science: a love story
49mPython for High Performance and Scientific Computing
06mPython Data Scientists 10x Faster
37mNipype