This branch contains the source for a Quarto-based portfolio of all coursework completed for DATA 624: Predictive Analytics at CUNY SPS, Summer 2026, by Group 3.
The portfolio covers time series forecasting, regression modeling, and model evaluation across two homework assignments and two group projects. It is published as a static site via GitHub Pages.
Group 3: Masoud Mahdisoltani, Andre Thomson, Desiree Thomas, Kevin Martin, Khandker M Qaiduzzaman
| Section | Description |
|---|---|
| Homework 1 | Time series decomposition, ETS, ARIMA (FPP3); predictor transformations and EDA (KJ) |
| Homework 2 | Regularization, tree-based models, Cubist (KJ); market basket analysis with association rules |
| Project 1 | Demand forecasting for Lone Star Appliances and hierarchical patient demand for Texas HealthLink |
| Project 2 | Regression modeling to predict beverage pH; champion model Cubist; Python vs. R comparison |
All data files go in resources/. Then from the repo root:
quarto renderOutput is written to docs/, which GitHub Pages serves.
The Python sections in Project 1 and Project 2 use reticulate. Project 2 caches slow grid search results to resources/py_results.csv and resources/py_preds.csv -- delete those files to refit from scratch.
In repo Settings, set Pages source to branch portfolio, folder /docs.