Predicting customer churn rates is critical for online education companies to increase customer loyalty and reduce customer churn. The customer abandonment prediction model helps take proactive measures by determining in which situations customers tend to abandon the service.
Dataset Description
The data set we created for this project includes customer demographic information, subscription types, payment methods, and customer behavior. The data set is contaminated with incomplete, erroneous and outlier data, just like real-world data.
CustomerID: Unique ID of the customer
Gender: Customer's gender (Male, Female, Other)
Age: Customer's age
Region: The region where the customer lives (North, South, East, West)
Tenure: Duration of customer stay in the company (in months)
SubscriptionType: Subscription type (Basic, Standard, Premium)
MonthlyCharges: Monthly charge
TotalCharges: Total paid charges
NumOfProducts: Number of products the customer has
ContractType: Contract type (Month-to-month, One year, Two year)
PaymentMethod: Payment method (Electronic check, Mailed check, Bank transfer, Credit card)
Churn: Customer abandonment status (Yes, No)
Project Steps
Data Loading and First Look: We will load the dataset and take a general look. We will examine the structure of the data and identify missing or incorrect data.
Cleaning of Missing and Erroneous Data: We will fill in the missing values and correct the incorrect data. In this step, we will perform the necessary cleaning operations to ensure the accuracy and integrity of the data.
Feature Engineering: We will create new features using the features in the data set and make the necessary transformations to improve the performance of the model.
Model Setup and Training: We will create a churn prediction model using machine learning models and train the model.
Model Performance and Evaluation: We will use metrics such as accuracy, precision, recall, and F1 score to evaluate the performance of the model.
Visualization and Interpretation of Results: We will present the analysis results by visualizing them and interpret the results obtained.
To solve, deeply examine and learn dozens of CRM analytics projects like this, you can register now for our 4-week, completely live and project-based CRM Analytics training.
Comments