In light of recent events like the government shutdown and rise of societal issues like climate change, we thought it was interesting how there are multiple opinions out in the open, yet we only hear about how our families, peers, and communities feel about those certain topics. Plus, news sources and google searches filter out certain views due to bias (take FOX news for an example), preventing us from forming our own opinions. Thus, using a widely used social media platform that solely depends on people tweeting their thoughts and opinions, we decided to focus on real time analysis of tweets to determine sentiment associated with events and societal issues in hopes to expand ours and others horizons.

What it does

We designed an app called Peak a Tweet that allows users to fuel their curiosity for a certain topic by searching for it as either a hashtag or user handle. Our app then takes the most recent tweets that mention the hashtag or user, analyzes them, and determines the general sentiment of those tweets. It relays back results in the form of emoticons corresponding to the attitude of the tweets.

How we built it

Using Swift, Sketch, TwitterKit, Twitter APIs, and CoreML modules, we built this IOS app. It uses Twitter Developer APIs and TwitterKit to gather data on tweets pertaining to a certain topic, gathers pertinent data from a JSON format, runs them through our machine learning module created from scratch using CoreML, and outputs a sentiment analysis about the attitude within the latest tweets.

Challenges we ran into

One of the main challenges we ran into throughout the coding process was creating the machine learning model from scratch. Apple only came out with this technology very recently (June 2018), so it was difficult to find help and resources to create this. We were able to create a simple module in Playgrounds and transfer it to our app. Another issue was figuring out how to incorporate user interface designs created in sketch into our Swift code. We had never experimented with Sketch, but were determined to have an appealing user interface design, as it is a crucial part of app development and use.

Accomplishments that we're proud of

We were able to use some of the latest and popular technologies such as machine learning in our app. We are proud of the amount of effort we put into learning how to create and incorporate our user interface designs into the code. We are also very proud of the idea we came up with, as it can be used in multiple different ways and can be used by anyone ranging from large corporate companies to curious individuals. We are also proud of completing this project in a short period of time!

What we learned

Through this experience, we gained experience in coding in Swift as well as working as a Twitter Developer with their APIs.

What's next for Peak a Tweet

In the future, our prototype can be further improved to be implemented across various other social media platforms like Facebook and Instagram. By gathering and filtering through a larger amount of data from multiple sources, our app can create a more accurate sentiment. We can also allow it to gather more than 100 tweets at a time, which can be done with premium Twitter APIs.

Built With

Share this project: