• 1-minute read

Time Series Data and Causality

This graph is great for demonstrating the danger of inferring causality from time series data alone.

A graph showing births and deaths in China from 2012. There is a sudden drop in births after 2016, when one-child policy was ended.
Taken from https://www.axios.com/2022/10/15/china-party-congress-2022-xi-jinping-third-term.

Keep reading

p-values

Adversarial Attacks on Linear Models

Don't Teach Statistics in High School