Probably Overthinking It is a venue for my articles on data science and Bayesian statistics. The articles are meant to demonstrate the tools of exploratory data analysis and visualization, and how they are implemented in Python. My goal is to show students and external audiences a process for using data to answer questions and guide decision making under uncertainty. Many articles use Bayesian statistics, which is an increasingly popular topic, but one that many people find hard to approach. I provide simple examples to help readers understand the important ideas and learn to apply the methods to other problems.
Using data from the General Social Survey, I plot religious affiliation by decade of birth. Among young adults, people with no religious affiliation will soon outnumber Catholics and Protestants.
Women in the United States have been marrying later for several generations. If current patterns continue, 30-40% of women born after 1980 will never marry.