Customer Churn Prediction

Project: Predicting telecom customer churn using machine learning

Tools & Technologies: Python, Scikit-learn, Pandas, Streamlit

  • Performed EDA and data cleaning
  • Used Random Forest feature importance to select top 4 features
  • Trained model on MonthlyCharges, Tenure, TotalCharges, Contract
  • Deployed prediction tool using Streamlit

Why is it Important to Know Customer Churn?

Customer churn directly impacts revenue and growth. On average, acquiring a new customer costs 5–7 times more than retaining an existing one. Studies show that even a 5% increase in retention can boost profits by 25–95%. Globally, telecom churn rates hover around 15–25%, while subscription businesses see an average of 6–8% monthly churn.



Glimpse of Code:

import pandas as pd
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
import joblib

# Data preprocessing and training steps...

Live App: Try the Churn Calculator