Do you feel that your Facebook News Feed is messy with countless unuseful news. Everyday, you have to spent hours and hours to read and scroll the News Feed to select only a few of them, which is not only annoying but also a waste of time. As a hacker, we see this annoying problem and it is an urge to solve it. That is the story of MyFeed, a tool to help you filter Facebook News Feed.

What it does

MyFeed only selects useful news and statuses. We analyse your timeline in the past to generate recommended keywords automatically. Since then, this would be the material for filter of News Feed in the upcoming news.

How we built it

We apply some techniques in Natural Language Processing to analyse old statuses. Firstly, a candidate keywords must be satisfied with 3 constraints:

  • number of minimum characters in each word
  • number of minimum occurrence of the phrase
  • number of words in the phrase.

2 main features, which are length and frequency of phrase/word. After that, they have to pass a few heuristic to be an acceptable keywords. Finally, is applied to chose select keywords.

Challenges we ran into

The lastest Facebook API doesn't allow to access home content of user. We have to downgrade the Facebook API. In backend, processing Facebook content is a huge challenge due to the complicated data such as images, videos, links, etc.

What's next for 005 - MyFeed

In the future, we will apply the state of the art in machine learning to recommend more accurate keywords for users. In addition, we also have a plan improve the user experience as well as more useful features.

In the next version, MyFeed will support iOS, Android, and many other popular social networks, i.e: Twitter, Tumblr, Instagram

Share this project: