Challenge Submission:

Ranked list: Design Doc: Algorithm Results: Frontend:


We are fortunate to have a vast array of news sources and aggregators that cater to our particular demographic, geographic, and sociopolitical identities. Even so, we should recognize that our favorite news websites can easily become echo chambers for beliefs that we already hold. In the United States and around the world, knowing the latest and most accurate information that promotes democratic and user-controlled consumption is ever so important. Therefore we built Newsly, a personalized news aggregator that allows you to control what you want to read and allows you to make the most informed and uncoerced decisions. Our current product revolves around news on Thailand, which we believe is a particularly apt use case given the social and political unrest and movements in recent months.

What it does

What many users fail to realize in the process of their daily news consumption is that the articles we are presented on a day-to-day basis is spoon-fed to us. We don't choose the factors that play into the order our articles appear in, the sentiment of the piece, or even the topics themselves (even though it may seem like we have that control). Newsly allows users to customize their feeds based on several factors we strongly believe are important to democratic consumption, such as sentiment analysis, peer reputation, news citations, political bias, Twitter article engagement, social bot promotion ratios, and even advertising aggressiveness.

How we built it

Newsly was built with a Postgres database, Flask server, and HTML/CSS frontend. We draw on several open source resources as well, including but not limited to Media Cloud, MediaRank, Twitter, and PyTorch.

Challenges we're submitting to

NewsQ for Social Good

What's next for Newsly

Newsly currently supports news feeds revolving around Thailand -- in the future Newsly hopes to democratize the world of news globally and allow for greater personalization of your news.

Share this project: