Overview
The entered data was from digitalized newspapers that were then uploaded to the Library of Congress. By including that existing data, they were able to create algorithms, n-gram shingling and locality sensitive hashing (LSH), on near-duplicate webpage
detection tasks. The complete data includes interactive maps and search engines that viewers can use to look up specific term clusters and see virality spreads.
Description and analysis of critical step
The most critical step in this project is objectively the development and the act of employing computational linguistics tools to automatically detect and align reprinted passages within vast archives of historical newspapers.
Link to tools and data
With the nature of this project, the connection between tools and data is very similar. Both are computational media that is entered into an algorithm. From that algorithm comes the data separated into clusters that contain identical phrases as organized data.
Documentation
Once I got used to how the website is set up, finding what I needed to analyze the methods of the project was not as difficult as it was before.
The language they use is generally easy to understand for someone with no computer analytics knowledge.
