xinjiehu76/cis5190-25f-projectb-8k5-combined

CIS 5190 Project B – 8.5K Combined News URL Dataset

Dataset Description

This dataset was created for CIS 5190 (Applied Machine Learning), Fall 2025, Project B at the University of Pennsylvania.

The dataset is designed for a news source classification task. It consists of 8,500 samples, where each sample contains a single URL corresponding to a news article published by a specific news outlet (e.g., Fox News, NBC News).

Each URL is treated as raw textual input.

Data Collection

The dataset was collected via web scraping of publicly accessible news websites.

• —Data collection and scraping : Chih Yu Tsai

Data Format

The dataset is provided as a single Excel file (85k_combined.xlsx).

It contains one primary column:

• —url: the raw URL string of a news article.

The classification label (news source) is derived from the domain name of the URL (e.g., foxnews.com, nbcnews.com) in project.

Intended Use

This dataset is intended for educational and research purposes, specifically for training and evaluating machine learning models for news source classification in CIS 5190.

Limitations

• —The dataset contains only URL-level information and does not include full article text.
• —The dataset may reflect biases present in the original news sources.