McNulty

For our third project, McNulty, my group chose to work with a bank marketing campaign dataset (1 of 3 pre-selected options we had to choose from) to predict who would say yes to a term deposit. To do this we first built a MySQL database on a DigitalOcean cloud server to store the data.

Next, we trained and tested multiple types of supervised learning classifiers (Logistic Regression, Decision Trees, KNN, SVM, Random Forest) to predict if a person would subscribe to a term deposit. Please see my blog posts for more details on the specific algorithms. In our case, logistic regression performed the best.

Finally, each person in our group had to come up with a different d3 visualization for the data. I choose to look at the group of people who said yes to the term deposit broken down by various demographics.

You can click on each of the groups and zoom in to see the specific count of each individual section. For example, clicking on the Marital Status bubble and hover over the word Single, you'll see that out of the 5,289 people who said yes to a term deposit, 1,912 of them were single.

Thanks to d3 guru, Mike Bostock, for the code.

Enjoy!