This repository contains materials for the MSCA 31008 Data Mining Principles Team Project.

Professor: Dr. Utku Pamuksuz

Team:

• Alicia Zhang: [email protected]

• Prafulla Ranjan Dash: [email protected]

• Serena Hanwen Xu: [email protected]

• Oleksiy Anokhin: [email protected]

• Surendiran Rangaraj: [email protected]

The structure of the repository:

• Data (raw and cleaned)

• Notebooks (9 Jupyter notebooks and 9 HTML files)

  • 1 notebook for data cleaning

  • 3 notebooks for exploratory data analysis

  • 3 notebooks for supervised learning (regression and classification)

  • 1 notebook for unsupervised learning (clustering)

  • 1 notebook for a recommender system

• Slides and Files (Project proposal and final presentation)

Project description

In this project our team applied the majority of algorithms we learned in the Data Mining class with Professor U. Pamuksuz.

In our work we used Starbucks App Customer Rewards Program dataset, which simulated about 140000 transactions. This data allowed us to ask multiple buisness questions about Starbucks customers and transactions and answer them, using different techniques (below).

Among them:

• Unsupervised lerning techniques (k-means, DBSCAN, hierarchical clustering)

• Dimensionality reduction techniques (PCA, t-SNE)

• Supervised learning techniques (regression, classification)

  • Decision Tree

  • k-nearest neighbors

  • Support Vector Machines

  • Random Forest

  • Gradient Boosting

  • AdaBoost

Please do not hesitate to contact us, if you have any questions.