RAYdar

Our banner

Inspiration

Our team realised that offensive messages are rampant in social media. Hence, we created a machine learning model to detect offensive messages against the LGBTQ+, different races and ethinicity and more groups.

What it does

It has 2 sub-features, a scanning feature and a typing check feature.

Scanning

When the user opens the extension menu, a pop-up wil show with a scan button, where the user can click the button in order to scan all the text inside the text input to see how offensive it is.

Typing check

When the user is typing in a messaging app such as discord or telegram, what they are typing is checked in order to check for any hateful words or phrases in what they typed. If what they typed is hateful, the app will send them an alert warning them that their message is hateful, and advise them to change their language to something more constructive.

How we built it

We used tensorflow to create the machine learning model in python. Then we used react to create the chrome extension. We used flask so that we can recieve the information from the model.

Challenges we ran into

a lot of challenge from using flask and running that.

Accomplishments that we're proud of

Created machine learning model to see how hateful some text is

Created flask server for backend to handle requests between the frontend extension and the ML model

Created browser extension which allows for background tracking of the user’s messages and easy checking of text on websites

What's next for RAYdar

Implementing it more deeply with messaging services such as whatsapp, telegram and discord. Also creating a desktop extension to use it outside of the chrome browsers.