It Just Works - Social Media Hate Detection - Hate Speech

• 

  An image of the website

Inspiration

I had a note tracker that I had previously coded, that I thought I could turn into a social media-like website. And, as the challenge was hate speech, I thought that I needed a way to detect and punish hate speech and its poster.

What it does

On the Tweeeter website, you can toggle between consumer and admin view. In consumer view, you can create multiple posts with custom titles and body texts. In admin view, you can see all flagged posts that possibly contain toxic messages. The admin then has the ability to punish the poster if the message is hateful.

How we built it

As said before, I had a previously coded note tracker that I coded for a different project, and I turned that into the Tweeeter website. The hate detection is determined by comparing the content of the posts to a list of hateful words and phrases. Everything was coded completely in p5js, and no CSVs were used. We used only one library, which was jQuery.

Challenges we ran into

I originally wanted the hate detection to be machine learning based on ml5, but I had never done machine learning before and I found I wouldn't have enough time to do so. So I instead resorted to a manually made list of hate words and phrases.

Accomplishments that we're proud of

Most of my coding projects have been done in Processing, and I rarely code websites. So this project is probably my biggest website yet. Therefore, I am proud of the entire project whether I win or not.

What we learned

Although I did not get to implement machine learning, I still learned a lot about ml5 while trying to implement it.

What's next for It Just Works - Social Media Hate Detection - Hate Speech

Next, I will probably try and implement the machine learning, as I will have more time and less pressure to do so.

Built With

• p5js