Inspiration
We thought that the constraints from NewsQ looked very stimulating and intriguing, and we have a passion to show a neutral side to an issue through news, so we decided to find a way to do so.
What it does
Our iOS app created by Xcode allows users to input links of news articles and then web scrapes the article itself as well as information about it such as its news source title. With that information, our app uses sentiment analysis to measure the bias of the article. Our app, from the information we collect from the analysis, will have another screen that implements data analysis and ML to display the list of news sources and articles based on the value calculated from the sentiment analysis algorithm. In essence, the users will then have an unbiased, computer-created analysis to more accurately judge the validity of the articles that they view.
How we built it
We built the app through Xcode and Swift, implementing Machine Learning and Web Scraping. We did extensive research into how to implement certain processes that we needed. We used github to share the code amongst teammates.
Challenges we ran into
We were unfamiliar with several elements of our app at first. We had minimal experience in CoreML, so it took time to understand how to use it properly with our JSON file. Additionally, working at the intersection of HTML files and an iOS app was particularly challenging as well, as working with converting those HTML files into a format for our model to interpret
Accomplishments that we're proud of
We are really proud that we could tie in two very different aspects of computer science together. Combining classifiation models and machine learning with the connections between HTML and website data and our iOS app was our proudest accomplishment.
What we learned
We learned a ton of information about ML, and even exploring through HTML files in iOS was quite informative.
What's next for Newsable
For the future of Newsable, we want to be able to implement a search algorithm for articles so that the user only needs to know the title of the article, and the app will extrapolate the hyperlink of the article. Additionally, we want to widen the range of our sentiment analysis, so as to more accurately predict the bias of an article. Or, we could even go so far as to create the article, but modified, so that it is written in an unbiased fasion.
Log in or sign up for Devpost to join the conversation.