How does the world really feel? With social media being a major presence in our lives, sometimes bigger than regular news, itself, our team set out to create a fun, unique app that can truly describe how areas around the world feel in response to a certain keyword, like 'brexit', 'summer', or 'trump'.
This is an application, currently for Android, that allows the user to type in a keyword. Upon submission, the user is brought to a map of the world, which is appropriately shaded for each region based on the sentiments of the tweets sent from that area. All tweets are analyzed with a machine learning algorithm that gives the tweets a coefficient of positivity and negativity, which we use to determine the shade of the specific region. Sample tweets hover around the regions, too, giving the user a very immersive and cool experience.
The front-end team (Julian Chow, Rishi Rabheru, Ashley See, and James Tavernor) used Java to code the user interface. The back-end team (Anthony Alridge, Nicholas Li, and Shravan Nageswaran) used Python to import and customize a machine learning algorithm to analyze the sentiment of the tweets, retrieve tweets from many regions, and combine these two to create a map that linked coordinates of the world to the appropriate colour of the map representation. Ultimately, we connected these programs well and made the full application.
The API would only provide us with 2500 tweets per minute (and we wanted a lot of data). Because of this, we limited the ICHACK version of the app to major cities in the United Kingdom and Ireland, but we look to continue developing the app to make it a worldwide representation.
We are very proud of optimizing the machine learning algorithm to process each tweet quickly. Additionally, implementing and manipulating a map in Android is quite challenging, and our front-end team did a phenomenal job of doing that. Plus, all of our code is readable, and the app, itself, is simple for users to pick up and enjoy!
We learned that processing data poses its challenges - especially when we want to analyze millions of tweets! Additionally, from ICHACK, we taught ourselves how to combine front-end and back-end work to accomplish the endeavor of a comprehensive application. It was a very enjoyable, hard-working, and rewarding experience.
Heatwave is continuing to be improved. Once we can process more data, we will make Heatwave assess all cities around the world. Additionally, we look to continue improving the graphics, adding a blend method that will appropriately blend two colours in adjacent regions to make the entire map look more cohesive. All in all, we wanted to create an app that is unique, insightful, and - quite frankly - fun, and that is what we did! But we are not done yet.