All Posts

Ridge, lasso and elastic net returned identical test error, to the last digit. The reason is a mistake hiding in a workflow you have copied.

Your model says 1.60. That cannot be a probability. Four things that go wrong reading logistic output, and none of them throw an error.

What happens when you bring talented people together with real problems to solve and the support to actually finish. The story of the Data Heroes Challenge and what it teaches about building data culture.

I removed 53 outliers. R-squared rose from 0.818 to 0.856 and the model got worse. Every number that improved ignored the rows I deleted.

Your slide says $368. A colleague who ran the same data says $422. Four ways a regression coefficient means less than it appears to.