The news is mostly negative, and Covid-19 hasn't made the situation better. We want to provide a source of real news that contains only the positive stories from news of the day.
What it does
It's a webapp that only shows you happy news.
How we built it
Our website uses two news APIs to get the news (Bing, New York Times). Once we get the news, we use Microsoft's Text Analytics to perform sentiment analysis on each of the headlines. If it's not a 'positive' headline (according to the sentiment analysis), we filter it out. Once all the unhappy stuff is filtered out, we upload the positive news stories to an Azure SQL database. We use a serverless Azure Function to automatically perform this news-gathering/filtering/uploading process once every hour, so that we always have fresh news. The user views the news on a Blazor web app, hosted at happynews.azurewebsites.net. When a user loads the page, instead of performing the news-gathering/sentiment-analysis process every time, it just retrieves the pre-filtered news stores from the SQL database, which cuts down on API usage and resource strain drastically.
Challenges we ran into
Getting the web app to access any kind of info is impossible without an API, so we had to make an API to translate between the database and the website.
Accomplishments that we're proud of
Getting this all to work.
What we learned
Sometimes AI can be wrong in terrifying/hilarious ways.
What's next for Happy News
More data sources and better UI!