LegitiFact Source Checker Bot

Visit the bot in action on the Legitifact Facebook Page.

Usage

perl ./run-source-checker [time]

The script pulls all link shares made by major Facebook news outlet pages, and reshares them while providing evaluations on their credibility, bias, and accuracy, backed by research from established polling and fact-checking institutions. The app takes an argument, $ARGV[0], the number of minutes of recent posts to query from each page.

Inspiration

In the context of the current media landscape, media literacy has become an invaluable necessity. The goal of this project is to educate social media users by evaluating news articles shared by widely viewed Facebook pages for accuracy, trustworthiness, and bias. We hope to promote greater media literacy among Facebook users in order to improve the spread of verifiable, trusted sources, and encourage readers to consume sources that present diverse viewpoints.

Function

Our Facebook page does two things.

Content Aggregator

First of all, it is a content aggregator. It continuously queries major news/media pages on Facebook, and reposts them onto its own page. It queries recent posts in near-real-time, from more than 50 major Facebook news media pages. This allows a diversity of media opinion to be presented equally alongside its peers.

Source Evaluations

Second, our bot presents a succinct but informative evaluation of each source linked from a major Facebook page. This evaluation informs the reader about the credibility of the source as seen by a diverse sample of respondents surveyed by the Pew Research Center. Additionally, it uses information from the study to acknowledge the existence of partisan differences in the perceived of each source. Lastly, it links readers to the Pew Research Center's 2014 research study on Political Polarization & Media Habits from which we obtained this information.

Build Details

This script is written in perl, taking advantage of the Facebook::Graph API to interface with Facebook Pages. The API is used to query and retrieve public posts made by major public Facebook pages. Recently created posts that contain links are extracted and compared with our existing database of news source evaluations from the Pew Research Center. These posts are then shared as new posts on the Legitifact public page, along with a brief summary evaluation of the source. By sharing the same post as the original source, we take advantage of Facebook's News Feed algorithm to guarantee that our post containing the evaluation is consolidated together with the original, so that users will be able to view our automated post at the same time as the original.

Challenges

Because of recent privacy related changes to the Facebook API, it was very difficult to access posts not made by the users. We got around that by posting onto our projects own page, and recognizing that Facebook's News Feed algorithm will automatically collect posts sharing the same link. Since we obtain our links from other Facebook pages, our source evaluations will always appear next to the original poster's page on a user's News Feed.

Expanding on the project

One of the unfortunate limitations of this project was the lack of peer-reviewed, academic research on the credibility of various internet news sources. In the Information Age, news media giants such as CNN compete with smaller "viral" news sources online. The vast majority of our data comes from a research study Pew Research Center carried out in 2014, targeted at news media in general, not on the internet specifically. I believe our application can be improved to be much better targeted at a Facebook news audience with more research data on the media credibility by source in the Information Age. PolitiFact Scorecard tracks the accuracy of a number of statements made by major news sources over a period of time, but we decided it was unsuitable for our project. Because of the nature of the website, Politifact intentionally cherry picks statements, and it does not claim to use peer-reviewed methodology to arrive at the Scorecard.

Share this project:
×

Updates