A simple application that based on a given training set makes predictions about the test set.
Application will require a sufficient amount of learning data in csv format. First column should contain target values - one which should be predicted. Other columns are treated as features of a training sample. With learning data provided you can choose which learning algorithms you want to use and start the process. Now you can pass new, not-yet-classified data and see some predictions based on them.
Application achived a top 5% score at kaggle Titanic challange. It is a good baseline for any ML classification problem.