Project

Fact checking visual analytics on Facebook newsfeed

In this work, we present a novel visual analytical tool, intended to improve the Facebook experience which produces an interactive visual analytic presentation of a user’s newsfeed to help understand bias trends existing among the user’s social group. The visual analytic tool provides steps to compare posts from different political spectrums and to validate the sources of the post in the Newsfeed. The tool also presents the propagation speed of the information seen. The spread of false information is prevalent on many different platforms of social media. Fake news and other invalid content are published for a variety of reasons like trolling, clickbait or smear an adversary. In this work, we focus on the Facebook platform. Facebook has a false information problem impacting its users. Depending on the user's social group the user may see a specific post based on the mutual interests of the group. Lots of news articles from different sources might be propagated, liked, and shared by anyone based on a sensationalized title. To develop our visual analytic tool to visualize the validity of posts based on the source from which the content was generated we collect and aggregate the Facebook feed data. We input the feed data to the tool provided by Graph API. The feed data is organized considering ID, content, number of shares, time of posting, associated media files and the source. The processed information is filtered to their respective categories and fields. Next, we statistically analyze the Facebook feed and input this data into a novel visualization which summarizes the different political spectrums seen in the posts. We designed and developed this visualization, using JavaScript, D3.js, Python and Graph API. We present a summary view which aggregates all the data, and the ability to zoom in to understand a specific post for its “genuine” factor. Data from every post is analyzed and classified into appropriate categories like Family, Health, Education, Entertainment, News, and, Digital. In this project, the emphasis is made on feeds related to Politics showcasing the five key factors on the post and finally representing the amount of truth in the post through visualization. We consider minute details in the post and apply suitable algorithms to visualize the data. Also, we consider the feedback from the user to provide more control and help the user be proactive about the information in their newsfeed.

Project (M.S., Computer Science)--California State University, Sacramento, 2018.

In this work, we present a novel visual analytical tool, intended to improve the Facebook experience which produces an interactive visual analytic presentation of a user’s newsfeed to help understand bias trends existing among the user’s social group. The visual analytic tool provides steps to compare posts from different political spectrums and to validate the sources of the post in the Newsfeed. The tool also presents the propagation speed of the information seen. The spread of false information is prevalent on many different platforms of social media. Fake news and other invalid content are published for a variety of reasons like trolling, clickbait or smear an adversary. In this work, we focus on the Facebook platform. Facebook has a false information problem impacting its users. Depending on the user's social group the user may see a specific post based on the mutual interests of the group. Lots of news articles from different sources might be propagated, liked, and shared by anyone based on a sensationalized title. To develop our visual analytic tool to visualize the validity of posts based on the source from which the content was generated we collect and aggregate the Facebook feed data. We input the feed data to the tool provided by Graph API. The feed data is organized considering ID, content, number of shares, time of posting, associated media files and the source. The processed information is filtered to their respective categories and fields. Next, we statistically analyze the Facebook feed and input this data into a novel visualization which summarizes the different political spectrums seen in the posts. We designed and developed this visualization, using JavaScript, D3.js, Python and Graph API. We present a summary view which aggregates all the data, and the ability to zoom in to understand a specific post for its “genuine” factor. Data from every post is analyzed and classified into appropriate categories like Family, Health, Education, Entertainment, News, and, Digital. In this project, the emphasis is made on feeds related to Politics showcasing the five key factors on the post and finally representing the amount of truth in the post through visualization. We consider minute details in the post and apply suitable algorithms to visualize the data. Also, we consider the feedback from the user to provide more control and help the user be proactive about the information in their newsfeed.

Relationships

Items