Forecasting the outflow of users of a telecom company

Online courses are becoming more and more popular. The online education market is constantly growing, especially a sharp jump occurred during the pandemic, which literally forced us to reconsider the attitude to educational processes completely. Online courses and lessons are gaining popularity every year, the audience of such courses is from schoolchildren to pensioners, respectively, and the requests of all users are different. Educational organizations strive to personalize communications as much as possible and predict user choices as accurately as possible in order to improve the customer experience.

To develop a model for predicting the most relevant educational course offer for the user based on Netology data.

Table of Contents

• General Info
• Technologies Used
• Project Status
• Room for Improvement

General Information

• Task: Predict the course of which direction the user will buy.
• The purpose of the project: Based on the data of user behavior on the site and his training in online service courses, determine his preferences in the field of online education, patterns of behavior of different groups and create a machine learning model to predict the possibility of a student buying a course and its orientation.
• The model will help the service to build a system of course recommendations for all users of the online service. It will allow you to optimize expenses and increase income by recommending courses tailored to individual interests.

Technologies Used

• Pandas - version 1.3.5
• CatBoost - version 1.0.6
• Numpy - version 1.21.6
• Sklearn - version 1.0.2
• Seaborn - version 0.11.2

Project Status

Project is: complete

Room for Improvement

To do:

• Additional study of user experience in different groups.
• In - depth feature Feature Engineering