written by Eric J. Ma on 2019-05-11 | tags: pycon software development sprint open source
Together with my colleague Zach Barry, we led a pyjanitor sprint! It was really fun, looking forward to more in the future!
written by Eric J. Ma on 2019-05-10 | tags: programming coding data science
Some realizations on how disruptive context switching can be. My first thoughts on how to deal with it during a sprint...
Read on... (240 words, approximately 2 minutes reading time)written by Eric J. Ma on 2019-05-10 | tags: pycon 2019 conferences data science
My real-time thoughts on SciPy 2019’s tutorial and conference days. It’s been a pretty awesome experience thus far, though shouldering three tutorials was indeed a marathon for me, much as I love teaching!
Read on... (383 words, approximately 2 minutes reading time)written by Eric J. Ma on 2019-04-29 | tags: pycon python data science conferences
I'm headed out to PyCon 2019! This year, I will be co-instructing two tutorials, one on network analysis and one on Bayesian statistics, and delivering one talk on Bayesian statistics.
The first... (read more)
(329 words, approximately 2 minutes reading time)written by Eric J. Ma on 2019-03-24 | tags: data science machine learning
Have you heard of variance explained as a loss function and machine learning metric? Turns out it’s quite useful and interpretable. I’d like to share this new learning with you.
Read on... (440 words, approximately 3 minutes reading time)written by Eric J. Ma on 2019-03-22 | tags: python hacks tips and tricks data science productivity coding
In praise of functools.partial, and how I used it in a Flask/Bokeh app!
written by Eric J. Ma on 2019-03-20 | tags: data science productivity
My tooling, routines, and techniques for getting things done and learning new things!
Read on... (825 words, approximately 5 minutes reading time)written by Eric J. Ma on 2019-03-01 | tags: data science programming best practices
In this Q&A-style blog post, I detail how data scientists can begin to engage in pair coding as a more common practice in our day-to-day work, and why we should spend the time to do it as much as we can afford.
Read on... (840 words, approximately 5 minutes reading time)written by Eric J. Ma on 2019-01-28 | tags: data science data products minimum viable products
I would like to encourage you to build more "minimum viable products" of your projects. Come learn why they’re so valuable!
Read on... (100 words, approximately 1 minute reading time)written by Eric J. Ma on 2019-01-21 | tags: scalability bayesian model dose response parameter learning model specification convergence shrinkage large dataset nuts mcmc advi variational inference neural networks random sampling biochemistry data modeling
I've been exploring a Bayesian hierarchical 4-parameter dose response model at work. Initially, I used a few thousand samples for prototyping, but I've now scaled up to 400K+ samples. Fitting the model with NUTS would've taken a week, but ADVI did the job in just 2.5 hours. 🚀 This experience has given me a new appreciation for ADVI, even in simpler models with large datasets. 🧠
Read on... (365 words, approximately 2 minutes reading time)