Healthy Messages

Our UI experience

Inspiration

According to a 2023 study, depression rates have increased by 70% in the last 25 years, with nearly one in 10 people reporting symptoms of depression. Depression, alongside many other mental health issues, has shown a strong linkage with social media usage, especially with teenagers. Social media contains many aspects that may cause damage to mental health, but one that has a significant impact and our team has experienced personally is targeted hateful content.

Hateful content on social media platforms manifests in various forms, ranging from discriminatory remarks and derogatory language to explicit acts of harassment and cyberbullying. Such content not only contributes to the deterioration of mental well-being but also fosters a toxic online environment that can have far-reaching consequences. As prevalent as this issue is, a solution needs to take place.

What it does

Our application utilizes our self-trained machine learning model to automatically identify hateful, negative, and oppressive content, and remove this content from various social media platforms. Users can prevent discriminatory comments on their posts and block others from sending hateful messages to them.

How we built it

For our machine-learning model, we combined datasets of twitter data and Dynamically Generated Hate Speech and used sci-kit-learn to develop our model. In the end, we reached a 90% model accuracy.

Challenges we ran into

Initially, our machine learning model had low accuracy, and required us to compile multiple datasets together to reach a usable state. Also, we had issues identifying and utilizing a reliable API for editing media content on Instagram and Youtube. Our UI was written on an unfamiliar framework, and we lacked reliable documentation to build our application with.

Accomplishments that we're proud of

With little experience in fine-tuning machine learning models to custom datasets, we learned how to develop a reliable machine learning model with a 90% accuracy. We also utilized many APIs and frameworks that we were otherwise unfamiliar with.

What we learned

Making Healthy Messages, we addressed a problem prevalent in all of our lives and created a reliable solution. We learned that through software and determination, real problems in our lives can be alleviated.

What's next for Healthy Messages

We plan to integrate Healthy Messages with other social media platforms besides Instagram and Youtube, and eventually create a mobile application that runs our filtration software in the background.