You have to compare your rule based derived label (Churned/Not Churned) with the actual label (provided) to find the difference/error. Exclude those customers which are in testing label file. Show the difference in percentage.

Task 1:

 Preprocess the dataset and make a feature set (provided above).

 Label the customer as churn/Not-churn?

 Make a model for churn prediction, i.e., provided customers in test data will churn or not?

A customer will churn if he/she has no visit in a week.

EDA:

 Calculate the week with the highest earning?

 Identify the most valued customer?

 Categorize the customers into 3 groups (i.e., Poor, Mediocre, Rich)

 Will we able to figure out churn important factors from available data?

 How Churn Rate changes with Historical Visits? Answer this query with the help of graph?

You have to submit the predicted outputs for customers provided in Submission.csv on Kaggle.