T9-Outlier-22FA

Project Topic: Reducing churn by identifying high-risk customers for the Telecom industry

About the Project:

• A major telecom company's postpaid business of voice-only plans is struggling to maintain its strong foothold in the local market due to:

▪ High churn rate amongst customers leading to a revenue decline of ~500k USD every month.

▪ The decline in overall customer base (high churn rate combined with low acquisition rate), leading to a decline in total market share.

• The project focuses on applying predictive analytics and identifying the factors that impact customer churn.

SMART QUESTIONS:

1. What are the factors which are driving customer churn?

2. Does credit rating have an impact on churn rate?

3. Does a high number of active subscribers lead to a low monthly revenue?

4. Which occupation category has high churn?

5. Do dropped calls and blocked calls have an effect on customer churn?

6. How many customers traveling outside the US are retaining the Telecom plan?

7. Does the monthly revenue average differ for different occupations?

Modeling Methods:

Logistic Regression, KNN, Decision Tree, Naïve Bayes, Random Forest

Link to our dataset:

https://www.kaggle.com/datasets/pranavchandaliya/telcom-churn-dataset