Many people who work with data are familiar with Principal Components Analysis (PCA): it’s a linear transformation technique that’s commonly used for dimension reduction, as well as for the orthogonalization of data prior to downstream modeling or analysis. In this article, we’ll talk about another PCA-style transformation: the sphering or […]

Estimated reading time: 11 minutes

Introduction A common question in analytics, statistics, and data science projects is: how much data do you need? This question actually has very specific and clear answers! A first good answer is “it is good to have a lot.” Let’s dig deeper and get some additional more detailed quantitative answers. […]

Estimated reading time: 10 minutes

Introduction The goal of this note is to try and characterize excess generalization error: how much worse your model works in production versus how well it appeared to work during training. The clarifying point is excess generalization error (also called overfit) isn’t so much the model performing unexpectedly poorly on […]

Estimated reading time: 13 minutes

Introduction I want to spend some time thinking out loud about linear regression. As a data science consultant and teacher I spend a lot of time using linear regression and teaching linear regression. I have found each of these pursuits can degenerate into mere doctrine or instructions. “do this,” “expect […]

Estimated reading time: 12 minutes

In most of our data science teaching (including our book Practical Data Science with R) we emphasize the deliberately easy problem of “exchangeable prediction.” We define exchangeable prediction as: given a series of observations with two distinguished classes of variables/observations denoted “x”s (denoting control variables, independent variables, experimental variables, or […]

Estimated reading time: 13 minutes