Project information
- Category: Machine Learning Classification
- Course: Machine Learning Zoomcamp
- Project date: 03 November, 2021
- Project Repo: https://github.com/AbdassalamAhmad/customer_subscribtion_prediction
- Project Live Demo: https://share.streamlit.io/abdassalamahmad/customer_subscription_prediction/main
Project Details
The project is about the prediction of whether potential bank clients will subscribe to the term deposit or not. Here are the main steps I did to complete this project:
- Data cleaning.
- EDA, visualizing the numerical and categorical features, feature engineering, checking for imbalance data.
- Training, validating and tuning multiple ML models parameters like: Logistic Regression, Decision Tree Classifier, Random Forest Classifier and XGBoost.
- Choosing the best model and deploy it locally using production ready server like waitress on windows and gunicorn on linux, and also deploying Docker container locally.
- Deploying the best model using flask to the cloud on PythonAnywhere for everyone to test.