Me and a team of friends from college (We are 2nd Year Students) recently created and deployed a webapp built to detect and flag fake Twitter Profiles using an AI Model trained on Profile Data. There is nothing much to say except that Random Forest Models were used and we also integrated Firebase Realtime Database to store a record of all profiles previously flagged by our webapp. The code itself is available at this GitHub Link and deployed at This Streamlit Link but due to recent changes to the Twitter API which migrated the endpoints our project used to the paid tier, it is since no longer functioning because it used API to gather data about the profile you were checking for. A demo video of how it previously worked is available here . Alternatively a seperate version of the webapp exists at this link where the model runs fine but the Parameters will have to be inputted manually instead of the API fetching them.
I learnt a lot about AI, Streamlit and Python in general and thoroughly enjoyed making it. It was built as our submission for the Problem Statement 1364 of the Smart India Hackathon 2023, and while we didn’t win, we did have so many new experiences.
Anyways, that’s all I had to share. I have other projects that I’ve built and deployed to streamlit too, although none as impressive. Anyone interested in discussing about these more may feel free to reach out.
Good Day!