Apollo

Apr 30, 2021

From September 2020 to April 2021, I worked on Apollo with Charmaine Wang, Eason Gao and Ryan Mok. Professor Alexander Wong was our project advisor. Apollo is our fourth-year design project. It is an extension for Google Chrome that helps users to identify false or misleading information on Twitter.

At the start of the project, we identified misinformation on social media as our problem space. We researched the problem space and explored different approaches to managing misinformation to identify an approach that we believed could actually address the problem. Ultimately, we decided to create a solution that presented additional information to help users identify false and misleading information. We identified three objectives for the project: assisting users in identifying misinformation, encouraging users to think critically and preventing users from spreading misinformation. We did not want to simply tell users if information was true as we did not want users to become dependent on our solution and we did not believe it was possible to create a fully objective automated fact-checker. As our timeline was limited, we decided to focus on creating a solution for desktop Twitter users.

Apollo was the result of our efforts. Apollo is a browser extension that presents additional information about articles posted on Twitter to help users fact-check. Apollo adds a button under each linked article in Twitter. When the button is clicked the Apollo sidebar opens and information about the article will load.

The final Apollo prototype presents three types of information in the sidebar: basic article information, article sentiment and related articles. The user can use this information to make an informed decision about the article quickly and easily.

When the user clicks the Apollo button to open the sidebar, a connection is established with the Apollo server. The server makes a request to get the article, parses the article and then processes the text to extract all the Apollo features. As a Chrome extension, the Apollo extension is written in JavaScript. The Apollo server is an Express server written in Node.js. For the related article system, I created a simple Flask service written in Python to utilize the PositionRank algorithm from the PyTextRank package. Datanews is used to search for relevant articles. The sentiment analysis feature uses Google’s Natural Language API. Apollo was run on the Google Cloud Platform App Engine.

I helped plan the project, analyze testing results and develop features for Apollo. The features I worked on included the related article system, the interface and the back-end server. The tasks I worked on were varied and included evaluating keyword extraction algorithms, tweaking styling and performing statistical tests on survey results.

Links to the Apollo code will be added soon.