Introduction Here you find a very simple, beginner-friendly data set. No sparse matrices, no fancy tools needed to understand what's going on. Just a couple of rows and columns. Super simple stuff. As explained below, this data set is used for a competition. As it turns out, this competition tends to reveal a common truth in data science: KISS - Keep It Simple Stupid What is so special about this data set is, given it's simplicity, it pays off to use "simple" classifiers as well. This year's competition was won by a C5.0 . Can you do better? Description We are looking at cold call results. Turns out, same salespeople called existing insurance customers up and tried to sell car insurance. What you have are details about the called customers. Their age, job, marital status, whether the have home insurance, a car loan, etc. As I said, super simple. What I would love to see is some of you applying some crazy XGBoost classifiers, which we can square off against some logistic regressions. It would be curious to see what comes out on top. Thank you for your time, I hope you enjoy using the data set. Acknowledgements Thanks goes to the Decision Science and Systems Chair of Technical University of Munich (TUM) for getting the data set from a real world company and making it available to be shared publicly. Also Vladimir Fux, who oversees the challenge associated with this data set. Inspiration This is a data set used for teaching entry level data mining skills at the TUM. Every year there is a competition as part of the curriculum of a particular course. This Data Mining Cup teaches some of the very fundamentals that are always worthy to be revisited, especially by pros abundant at Kaggle. For some of my thoughts see the verbose comments in the Kernel.