Jupyter Digest: 15 May 2017
TSFRESH, 100 days of algorithms, how JupyterHub tamed big science, colorizing photos.
- TSFRESH. TSFRESH is a “time series feature extraction based on scalable hypothesis tests.” In layman’s terms, it finds interesting things on a time-series chart for you automatically. The notebooks folder has Jupyter examples that show how to use it in your work, like this one that uses accelerometer data to figure out when you’re walking, climbing stairs, or just doing nothing at all. (Submitted anonymously.)
- 100 Days of Algorithms. Tomáš Bouda (@coells on GitHub) compiles a nice list of examples that illustrate a host of different algorithms with Python. If a title like “Day 14 – huffman codes.ipynb” lights you up, then you’re gonna love this (times 100). Also, bravo for not calling it “50 algorithms to whiteboard before you die.”
- How JupyterHub tamed big science: Experiences deploying Jupyter at a supercomputing center. If you’re trying to deploy Jupyter at scale in your org, then this session at the upcoming JupyterCon (Aug 22-23 in NYC) will be the place to be to learn how.
- Interactive Deep Colorization. Richard Zhang, a PhD candidate at UC Berkeley, uses this notebook to illustrate a technique he’s developed to colorize black and white photos (see below). Take that, Dorothea Lang.