The modern news environment is oversaturated, rife with misinformation, and driven by attention-grabbing clickbait content. While some platforms are striving to maintain journalistic integrity and excellence, the fact remains that there is simply too much information and coverage available for users to process. Local news outlets are also dying out at an alarming rate, meaning that coverage of the news that most closely affects people is becoming less comprehensive and less common.
As students, we sought a platform where our voices could be heard - a platform for productive discussion that works to solve issues rather than lament over and criticize them. Currently, discussions about local politics and news are largely limited to incensed rhetoric on minimally-read, densely-populated chat boards. Engagement feels futile, and a comment culture of malicious language and misinformation discourages participation.
To address this need, we created the Verum, a community-directed, town-hall-style news and discussion platform for micro-communities like our own Stanford community. With localized, exclusive groups, members are offered access to forums and conversations within communities of fellow students, co-workers, local residents, and others with shared characteristics.
What it does
The Verum promotes healthy, productive discourse about topics ranging from the broadest national issues to local legislation; it functions as a safe space for users to express their ideas and beliefs about issues most important to them and their individual communities without fear of being drowned out.
Optimized for discourse in microcommunities: The Verum offers small microcommunities that allows for all members to be heard. With high impressions per post, members are encouraged to post, and to post higher quality content.
Community-directed discussion: A machine-learning engine generates query results ranked by popularity and by the search history of the user within the site. It also integrates Microsoft Text Analytics Machine Learning API to sift out explicit language, hate speech, and malicious content.
Scalable backend construction: The Verum's backend database connection is built on Microsoft Azure, allowing for easy scalability and supporting future microcommunity development and propagation.
The Verum also features an Android application as well as a seamless mobile webview interface accessible on any mobile device type. With a versatile homepage layout, efficient posting functionality, and clean, contained menu sidebar, the Verum is just as easily usable on a mobile device as it is on a computer.
How we built it
The Verum was built using html, css, and js to handle the front end and UX with PHP and MySQL serving as the back-end connections. The server is hosted on Microsoft Azure for easy scalability and the application is built on the platform as a web app. We utilized Microsoft’s proprietary ML API for text analytics in our project to find keywords in our user’s reading history and suggest curated posts that would be of relevance. All of our team members helped contribute to the coding process and we split it up with half working on the front end and half implementing the backend. The project lasted the entire time at Treehacks and we hope to continue developing this idea more in the future.
Challenges we ran into
Building a dynamic web application on Azure was definitely challenging as this was our first time using any type of modern cloud hosting software. One of our members had experience with Mysql before but understanding how to set up Azure took some time. Moreover, our application has a really fleshed out backend design with all functionality like login/registration, dynamic post ranking, user posts, upvoting, and ML algorithms. These components all take a considerable amount of time when considering the effort it takes in orchestrating front and backend technologies. In particular, the Machine Learning implementation took a great deal of time since this was our first time working with the software. Using the Microsoft ML libraries was a really great help and we were ultimately able to implement our Text Analytics ML processing.
Accomplishments that we're proud of
We took inspiration from UI’s we thought were really innovative and combined it with our own creativity. The Verum has a clean and simple user interface and it took a lot of discussion and thought with our team to achieve what it ultimately looks like now. Moreover, a lot of the websites features are fully functional, which required a great deal of decomposition and programming to implement. We took it as a real challenge to build the backend functionality and we’re proud to say that all features that would be expected for a launch are implemented and working (login, registration, voting, posting, commenting, dynamic loading, inappropriate content filtering, infinite scroll, and mobile webview accessibility). Moreover, we were able to host on Azure, a cloud hosting platform that will allow the Verum the space and capability for rapid future scalability. We also developed machine learning applications to sort out hate speech and promote valuable opinions. This was particularly important to us because it allowed us to enhance healthy productive discourse on the Verum while applying some cutting edge technology to do so.
What we learned
This was our first time using Azure or any type of cloud hosting platform so we learned a lot about how these applications are built on the cloud and the ease of scalability. We were really excited to learn how to use Microsoft’s ML API and it taught us a lot about how ML applications are implemented in dynamic web apps. We were super excited by both products offered by Microsoft and will definitely use their applications even more extensively in the future. The project also taught us a lot of UI design and program decomposition. Being a platform with so many different functions that were crucial to implement for optimal UI and concept realization, we really developed our skills in isolating issues and collaborating to accelerate development of the Verum.
Developing the Verum has also forced us to confront some of the starkest realities of the modern journalistic sphere - that the human attention span is limited, and that quality of content and reporting is not users’ primary concern in news. We sought to learn why this has become the nature of modern online journalism, and how to combat it in favor of a productive and progressive news environment for the future.
What's next for Verum: Community Directed News and Discussion Platform
Content cleanup: We also hope to integrate Microsoft Text Analytics' Sentiment Analysis operations to improve the language processing engine accuracy and completely root out the potential presence of hate speech and malicious content on the Verum.
Algorithm refinement: We hope to refine our algorithm to rank posts based on the posting experience of individual users as well as on their inclusion of academic citations. This addition would give greater priority to posts that have in-text citations and references to other articles and posts, academic journals, and previous events. This would help promote a culture of educated and informed discourse, furthering the mission of the Verum.