Predicting Bank Customer Churn with Machine Learning in Python | Streamlit App + Power BI Dashboard

In this video, we take you step-by-step through the process of using Python and Machine Learning to predict bank customer churn. 🚀💳

Here’s what you’ll learn:
1️⃣ Churn Prediction Model:
• Built a machine learning model to predict whether a customer will churn and their probability of churning.
• Identified the most important factors driving churn using feature importance analysis.
2️⃣ Streamlit App Deployment:
• Created an interactive Streamlit app where users can input customer data to see:
🔹 If the customer is likely to churn.
🔹 The probability of churn.
3️⃣ Power BI Dashboard:
• Designed a detailed Power BI dashboard showcasing:
🔹 Insights from the model and analysis.
🔹 Predictions for existing customers.
🔹 Key trends and metrics for better decision-making.

Steps We Followed:
• Loaded raw data from Kaggle and formulated key questions.
• Preprocessed the data (cleaning, deduplication, scaling, and transformation).
• Visualized distributions and outliers to understand the data.
• Ran feature importance using Decision Trees and XGBoost Trees to identify the key drivers of churn.
• Developed and evaluated the model using an XGBoost classifier and a confusion matrix.
• Tuned the model for optimal performance with hyperparameter optimization.
• Saved the model and transformer using pickle and exported results to Excel for further analysis.
• Deployed the model in Streamlit and built a Power BI dashboard to present findings.

This project combines data science, machine learning, app development, and dashboard creation for a complete end-to-end solution to tackle customer churn.

👉 Watch now to see how we built it step by step and learn how to create similar solutions for your own projects!

#CustomerChurn #MachineLearning #Python #Streamlit #PowerBI #XGBoost #datascience

🔗 Chapters:
00:00 – Intro
02:11 – ML Process
03:26 – Raw Data Load
04:02 – Data Preprocessing
05:32 – Data Visualizations
11:08 – Running XGBoost/ML
14:04 – Oversampling Minority Class
16:10 – HyperParameter Tuning
18:48 – Storing the model & Results

Python Part 1 video:    • Predicting Bank Customer Churn with M...  
Streamlit Part 2:    • Видео  
Power BI Part 3:    • Видео  

Github Link: https://github.com/Pitsillides91/pyth...
Connect with me on LinkedIn:   / yiannis-pitsillides-8b103271  
Follow me on X: https://x.com/pitsillides91