Contains Multipage Streamlit applications showing all steps of machine learning pipeline with additional recommendations at the end.

With the enormous increase in the number of customers using telephone services, the marketing division for a telcom company wants to attract more new customers and avoid contract termination from existing customers. This churn prediction model would be able to provide clarity to the telcom company on how well it is retaining its existing custome…

This is an end-to-end AWS Cloud ETL project. This orchestration uses Apache Airflow on AWS EC2 as well as AWS Glue. It demonstrates how to build ETL pipeline that would perform data transform using Glue job/crawler as well as loading into a Redshift table. It also shows how to connect Amazon Athena to Glue Data Catalog, and Power BI to Redshift.

In this project, I aim to predict customer churn for Deutsche Bank using supervised machine learning. It involves data exploration, feature engineering, and building Naive Bayes, Decision Tree, Random Forest, and XGBoost models. Models are tuned, evaluated, and compared to identify the best approach for churn prediction.