Customer churn prediction is to measure why customers are leaving a business. In this I have looked at customer churn in telecom business.
I have collected the Dataset from Kaggle Website in which Customer account information- how long they’ve been a customer, contract, payment method, paperless billing, monthly charges, and total charges Demographic info about customers – gender, age range, and if they have partners and dependents. Services that each customer has signed up for – phone, multiple lines, internet, online security, online backup, device protection, tech support, and streaming TV and movies etc. given in thhe form of 7043*21 size data.
I have gone through several steps like Data Preprocessing, Data cleaning, Feature Engineering and Model building and I have then handle the Imbalance data using several techniques to improve the score of the minority and majority class such that i can improve the overall F1 score. With the help of Flask, HTML, JavaScript I have builded a web application for this.
Before balancing the Imbalance data-
After balancing the Imbalance data-
- Prgramming Language - Python 3.8
- Libraries - Pandas, NumPy, Seaborn
- Deep Learning Framework - Tensorflow 2.3
- Web Technology - Flask, HTML5, CSS, JavaScript.