John Geer

I build things with statistics, code, and prose.

Just a dog with a mustache

Photo by ginnerobot


How to Analyze Your Predictable Data: Anomaly Detection

Noticing unexpected observations among an ocean of expected ones.
This post was featured on and @analyticbridge.

Customer Retention with Statistical Learning

Predicting if customers plan to leave and finding the best intervention

Using Data Science to Estimate Annual Cycles in Housing

Even in the housing market, some things are predictable.
This post was featured on

Better Estimation When Perfection Is Unlikely: A Bayesian Example

A simple formula that is hard to derive.

A Better Way to Rank a Huge Number of Things

Statistics & Crowd Sourced News


Probability Distribution Family Tree

Some distributions are just special cases of other distributions.

US Federal Outlays & Receipts

It's how they say spending and income.

Visualization of How the US Budget Changes

Received the "Deep Thought Badge" and Honorable Mention in Google & Eyebeam's Data Visualization Challenge.


TED Views Predictor

A predictive model that uses the title, topic, and publish date of TED talks to estimate the number of times they will be viewed.

Bayesian Link Aggregator

A website that ranks links based on votes. It aims to learn the most from each vote by selecting which articles visitors vote on and using a bayesian estimator.

Interesting hack for following a topic HN-style [...]

Hilary Mason (@hmason), Data Scientist in Residence at Accel

Genetic Prediction Algorithm

A program that evolves a prediction model for a given data set.


I have a Masters in Applied Statistics. My specialties include predictive analytics, data mining, and machine learning. I presently work as a data scientist for a tech company in the Raleigh / Durham area.

Feel free to contact me.