Machine Learning Projects

This repository contains applied machine learning notebooks that focus on customer analytics, predictive modeling, and unsupervised learning techniques. The projects demonstrate end-to-end workflows including data preparation, modeling, evaluation, and interpretation of results.

---

Contents

• Customer_Churn_Analysis.ipynb — Predicting customer churn using supervised machine learning
• Clustering_and_Recommendation_Systems.ipynb — Unsupervised learning and recommender systems

---

Notebook 1: Customer Churn Analysis

File: Customer_Churn_Analysis.ipynb

Goal:
Analyze customer behavior and build predictive models to identify customers at risk of churn.

Focus:

• Understanding drivers of customer churn
• Predicting churn probability using supervised learning
• Evaluating model performance with appropriate metrics

What’s inside:

• Data cleaning and feature engineering
• Exploratory data analysis (EDA)
• Classification models (e.g. Logistic Regression, tree-based models)
• Model evaluation using metrics such as accuracy, precision, recall, ROC, and AUC
• Interpretation of model results and business implications

Skills demonstrated:

• Supervised machine learning
• Classification modeling and evaluation
• Translating model outputs into actionable insights
• Structured analytical reasoning

---

Notebook 2: Clustering and Recommendation Systems

File: Clustering_and_Recommendation_Systems.ipynb

Goal:
Apply unsupervised learning techniques to identify patterns in user behavior and build recommendation logic.

Focus:

• Customer / user segmentation using clustering
• Similarity-based recommendation approaches
• Hybrid recommendation concepts

What’s inside:

• Clustering methods (e.g. k-means or similar approaches)
• Feature representation for users and items
• User-based and item-based similarity using cosine similarity
• Hybrid recommendation logic combining multiple signals

Skills demonstrated:

• Unsupervised machine learning
• Clustering and similarity-based modeling
• Recommendation system fundamentals
• Practical application of distance and similarity metrics

---

Tech Stack

• Python
• NumPy
• pandas
• scikit-learn
• matplotlib / seaborn

---

How to Run

1. Create a Python environment and install the required dependencies.
2. Open the notebooks in Jupyter or VS Code.
3. Run each notebook top-down.

---

Author

Devin Mavric