Jose Unpingco, PhD

Although I'm quite convinced that I'll always be a writer first and a conference speaker second (or third or tenth), decided to record a quick cut of the same slides I gave at QCon a few weeks ago. youtube.com/watch?v=vkfzpQ…

At long last, the third edition of Python for probability, Statistics, and Machine Learning has been released! link.springer.com/book/10.1007/9… Overall, the book is now about one-third larger than the second edition!

@michaelxpettis Could you please explain why low household shares of GDP lower unit labor costs?

7/9 These low household shares of GDP work partly by lowering unit labor costs for manufacturers and partly as the obverse of the high shares of GDP plowed into logistics, infrastructure and direct and indirect subsidies for manufacturing.

1/9 Xi Jinping said this week that China must address and resolve the problems of "unbalanced and inadequate development" over the next five years. Among other problems he cites "low quality growth" and "weak innovation". reuters.com/world/china/ch…

Slides from my tutorial on machine-learning with missing values slideshare.net/GaelVaroquaux/… The tutorial explores the links and differences with classical missing-value practice in statistics: prediction calls for different tradeoffs. It gives theoretical results and practical tips

Excited to give a tutorial on data processing with missing values at #odsceast @odsc, jointly with @JulieJosseStat, starting now. I'll be focusing on supervised-learning with missing values.

Check out our new paper: Multiple Imputation with Denoising Autoencoder using Metamorphic Truth and Imputation Feedback arxiv.org/abs/2002.08338 #Datascience #MachineLearning

.@shelleylyford takes the podium, says @GMWFoundation has awarded over 545 grants totaling more than $213M to support initiatives that lower #HCcosts to enable #seniors to successfully #ageinplace

Thanks @ECMWF and @clavitolo for the opportunity to speak at the Research Reproducibility workshop. #repwork19 Here are my slides from the keynote: speakerdeck.com/willingc/scali…

Made a new post with notes and tips for developing with Cython! nicolas-hug.com/blog/cython_no…

I enjoyed making this screencast accelerating a processing pipeline several orders of magnitude with Numba, Dask, and CuPy/RAPIDS. youtube.com/watch?v=wANQkg… I did it all in a couple of hours. It was a blast to use everything together.

github.com/unpingco/Pytho… updated with conda and docker setup instructions to run the book's Jupyter notebooks. #Python #MachineLearning #Statistics

#IPython still have plenty of potential features. Even in CLI. Next in line support for mimetypes handlers. That is to say for example, display inline image (if terminal support it) or math, or call an external file viewer.

The 2019.07 release of Anaconda Distribution - focused on #Conda performance - is available for download. Install times have decreased by over half for large packages like PyTorch. bit.ly/2SLEeRL #Python #DataScience

Great blog Copula post by @twiecki twiecki.io/blog/2018/05/0…

A novel way @westhealth is looking at segmentation- particularly for organizations taking on financial risk and deciding who to target for enhanced services #geriatrics #healthcare #academyhealth #eGEMS #data egems.academyhealth.org/articles/10.53…

Second updated and revised edition of my Python for Probablity, Statistics, and Machine Learning has just been released by Springer. If you like your Python with a heavy helping of maths, then check it out.

Here are the slides and notebooks for my presentation on scikit-learn histogram-based Gradient Boosted Trees: slides docs.google.com/presentation/d… / notebook nbviewer.jupyter.org/github/ogrisel… #EuroPython #EuroPython2019 /cc @europython