SocialFly

Inspiration

Small businesses, fledgling non-profits, and teams trying to change the world are constantly working. They're building a product, forging innovation. Currently, social media is essential for the success of these groups, but to be focused on twitter and instagram when you have to upheave an industry, seems unimportant. SocialFly is focused on giving teams wings to use social media to their advantage, while not waisting time. More innovating, less counting followers.

What it does

SocialFly allows users to see specific Twitter data based on a query and analyzed for sentiment analysis by a neural network. It does the heavy lifting of a social media team, finding patterns in positive and negative responses towards possibly someone's product, a competitor, or an important social issue.

How we built it

We started the challenge with focus on user interface - the user experience should be easy, quick, and simple. A heavy emphases was put on the design of initial landing page, allowing users to simply input a search query and get results in their dashboard. In order to design the frontend, we used Next.JS (React), Tailwind, and Chart.JS. The backend was built using Node.JS and Express.JS, which queries the Twitter API, and send this data to our TensorFlow neural network for sentiment analysis.

Challenges we ran into

We really had a hard time implementing a python trained neural network into a Node.JS environment, but with a lot of work and dedication, we were able to create a small python API to use the saved TensorFlow model. We also put a lot of emphasis on user experience with the graphs, making sure the data would be presentable and easy for anyone, with any experience, to understand.

Accomplishments that we're proud of

We're really proud of developing a neural network that can find the sentiment of a tweet with an accuracy of almost 95%. This was a difficult task, but something we knew was essential to an app that could help those trying to learn patterns and trends in their Twitter query.

What we learned

We learned a lot about working in a team and developing a product not in a linear fashion, but developing with multiple people working on a project all at once. The machine learning process was also very informative, since though we've worked a lot on machine learning before, the deployment and integration process was new.

What's next for SocialFly

We mean to increase the amount of data displayed to users by using the premium Twitter API and deploying our application. We were limited by time and certain features could definitely be improved, for example our key words. However, this experience has made us so passionate about our project and we'll all be continuing our work on SocialFly later on after this competition.

Built With

• chart.js
• express.js
• flask
• next.js
• node.js
• python
• tensorflow