I help researchers in the social and natural sciences with data analysis, especially problems where a Bayesian approach might be useful. I work in Python or R, using PyMC, Stan, brms or whatever fits your project.

Bayesian methods are often a good fit when:

• your samples are small, or some groups have much less data than others;
• your data are hierarchical: students in schools, patients in hospitals, measurements on sites;
• you need to combine evidence from several sources, or include what is already known;
• you have missing data, measurement error, or censored observations (survival and time-to-event data);
• you want estimates with honest uncertainty, or need to make a decision based on them;
• you are working with survey data and need estimates for subgroups or small areas.

Free one-hour assessment. Each month I take a few free one-hour sessions. Tell me about your question, your data and what you have tried. We’ll meet for an hour, and I’ll tell you honestly whether a Bayesian approach would help and what the work might look like. No obligation; sometimes the answer is a simpler method you can do yourself.

For larger projects, I can bring in my colleagues at PyMC Labs.

About me. I’m the author of Think Bayes, Think Stats and Think Python, a professor emeritus at Olin College, and Principal Data Scientist at PyMC Labs. More about me.

Tell me about your problem

I read every message and reply within a few days. Your message goes only to me.