Bank Customer Churn Analysis

Analyzed banking customer data to identify patterns associated with customer churn. Performed data cleaning, Power Transformation for Age, outlier treatment, categorical encoding, feature engineering and scaling. Compared multiple machine learning algorithms including Logistic Regression, Decision Tree, Random Forest, KNN, SVM, Gradient Boosting and XGBoost. Built an Artificial Neural Network using TensorFlow/Keras and explored hyperparameter tuning using Keras Tuner. Gradient Boosting achieved 86.85% accuracy, while the tuned ANN achieved approximately 87.4% validation accuracy during tuning.

Telecom Customer Churn Prediction & Power BI Dashboard

Worked on the project end-to-end, including data cleaning, validation, transformation and exploratory data analysis. Built and evaluated a customer churn prediction workflow using Python and Scikit-learn. Developed an interactive Power BI dashboard using Power Query and DAX to analyze churn trends, customer segments and KPIs. Also developed a Streamlit application for churn prediction and prepared reusable model, encoder, scaler and feature artifacts for deployment.

NLP Emotion Detection using Machine Learning

Developed an end-to-end NLP classification project to detect emotions from text. Performed text cleaning and preprocessing using Regex and NLTK, and created text representations using Bag of Words, TF-IDF and Word2Vec. Compared machine learning models including Logistic Regression, Decision Tree and Random Forest. Logistic Regression with Bag of Words achieved 85.60% accuracy. Also developed a Streamlit application for model inference.