MTA Subway and DE-HNN Project

Description

This project combines two assignments:

1. MTA Subway Analysis: Leveraging MTA subway ridership data to analyze patterns, such as top stations and routes by ridership.
2. DE-HNN Reimplementation: Reimplementing the DE-HNN (Directed Edge Hypergraph Neural Network) model for chip design congestion modeling.

The project uses datasets and tools to explore graph-based models and deep learning techniques in real-world applications.

---

Getting Started

Prerequisites

Ensure you have Python (>=3.8) installed. Required Python libraries include:

• torch
• torch_geometric
• numpy
• pandas
• matplotlib
• seaborn
• scipy
• scikit-learn

For a complete list, see the Requirements section.

---

Datasets

MTA Subway Data

• GitHub Repository: MTA Subway Project
• Data Files: MTA Data Files

DE-HNN Data

• GitHub Repository: DE-HNN Project
• Data Files: Chip Design Dataset

Download and place the data files in the data folder before running either notebook.

---

Installation

1. Clone the repository:

   git clone https://github.com/okinahiru/DSC180A-Q1.git
   cd DSC180A-Q1
   

2. Create a virtual environment and activate it:

   python3 -m venv env
   source env/bin/activate
   

3. Install the required dependencies:

   pip install -r requirements.txt
   

Results

MTA Subway Analysis

The MTA Subway Analysis provided a deeper understanding into the intricacies of graph's and their role in machine learning.

Visualizations (saved in /plots/):

• Graph of subset of ridership data

---

DE-HNN Reimplementation

The DE-HNN (Directed Edge Hypergraph Neural Network) model reimplementation yielded:

• Model Performance:
  • Mean Absolute Error (MAE): 2.63

Visualizations (saved in /plots/):

• Graph plots utilized in EDA

---

These results showcased my journey through learning and understanding the intricacies of GNN and the DE-HNN model we will be working from in Q2