Losing employees is a very expensive cost for any employer. This project classifies employees who will leave based on survey data including satisfaction level, last evaluation, number of projects they have, average monthly hours they're working, time they spend at the company, if they got promoted in the last 5 years and a few other data points. The dataset has 15k rows and 10 features. Thus hopefully employers can work to try and reduce losing employees they don't want to lose.
After building a classifier, I stored the data in SQL and then visualize the relationships between how they work and why they leave. I found 6 important feature to predict churn. You can also go online here and see a live demo where you can check how it'd score for your own data.
If you want more information, please also visit this presentation.