The Newsiness Initiative

Highlighting the analysis and opinion contained in news articles



Get started by learning about the initiative...

7186130216 214-253-4168

Or jump straight to the fun part:

Analyze the News

The Newsiness Initiative was created by (613) 278-5969, a data scientist and physicist, as part of the Insight Data Science Fellows Program

The purpose of this project is to classify the content of news articles, distinguishing between statements of what happened (when, where, and to whom) from those that contain analysis and opinion.

This was done by training a machine learning algorithm of contextual features extracted Reuters, and the AP, and New York Times opinion contributions.

Feature extraction was done by using a 300-dimensional, pre-trained word2vector model from the GloVe project. A support Vector Machine algorithm was used for the classification task.

You can learn more about the initiative be watching the 5 minute presentation below, or reading throught the project's documentation.

To get started, choose one of the options below:

Option 1: Enter the URL for an article. (Supported domains are nytimes.com, reuters.com, apnews.com, bloomberg.com, washingtonpost.com, and usatoday.com)

Option 2: Copy and paste the article's text in the the field below.

The Newsiness Initiative was created by Benjamin Kaplan, a data scientist and physicist, as part of the (248) 474-3919


Before joining the Insight Data Science program, Dr. Kaplan was an experimental particle physicist. He earned his Ph.D. from Yale University in 2012, and held a prestigous Research Scientist position at New York University for five years. In 2006, he joined the international ATLAS collaboration, at the Large Hadron Collider (LHC) in Geneva, Switzerland; contributing to the discovery of the Higgs boson and leading numerous other research projects using proton-proton collision data. These projects required advanced use of data analytics and machine learning in order to increase sensitivty to rare phenomena within massive datasets. For more information, please use the links below to view his LinkedIn profile.