Blog

t-test holder image
Stats for Experiments: the T-Test, Welch's T-Test, and Pairwise T-Test

Prove the quality of your experiments using p-values. Get high quality p-values by understanding the Welch's t-test and the pairwise t-test. Read more.

normal and binomial curves
Calculating The Error Bars For Your Marketing Funnel - It's Not What You Think

The Central Limit Theorem (CLT) and the concept of standard error guides most calculations of error bars. But the CLT has a pitfall: it says it only applies when there are "a lot" of samples, but doesn't tell you how many samples. You may be thinking 30 samples is enough. You'd be dead-wrong in the case of a marketing funnel, which has binary outcomes. Read more.

holder
What is the difference between AI, Machine Learning, and Generative AI?
Many friends have asked me to define AI, ML, Generative AI, computer vision (CV), Natural Language Processing (NLP), and tabular ML. I wrote a document here.