Eric J Ma's Website

If df.iterrows() is slow, what then is the alternative? Read on to figure out how to make looping over dataframes 1000X faster :).

I test-drove the Python package pandera this month, and I like its usage paradigm. Come read on to learn how you can incorporate pandera into your workflow!

Why do software skills matter for data scientists? We might have heard that it matters for our workflow, but what about for organizing knowledge? In this essay, I argue that practicing good software skills has those benefits and more.

In this post, I detail some of my thoughts about the use of "data challenges" for hiring data scientists. Though I have only used it in hiring data science interns, I think some of the lessons I've learned from test-driving the process can apply more generally.

I like writing in notebooks, for the ability to quickly prototype. But can we treat Jupyter notebooks as scripts that we can execute? The answer is yes, and in this blog post, I'll show you a few of the simplest ways to do so.

I test-drove Hey, a new email product launched by Basecamp recently, and I'm ready to pay. Read on to find out why!

I came to the epiphany today that "statistical testing" protocols are nothing more than canned model comparisons (that sometimes have convoluted interpretations). Come read why!

I've been reflecting on the way I learned statistics, and I think I learned it in a flawed fashion. In response to this reflection, I've began reworking an introduction to Bayesian statistical inference that focuses on statistical story telling. Read on to find out more!

What gives us the logical leap from a joint distribution $P(A, B)$ in abstract, to the more tangible $P(H, D)$ (for hypothesis and data? Here's an excerpt from a tutorial in the making about Bayesian statistics.

How in the world did "inference" come to mean "prediction" amongst deep learners? I have no idea, but I do have a rant.