Florida shootings show us that there's fairly consistent data that these extremists that go out and commit mass murder tend to suffer from mental disabilities and an anger against society, wether that be their workplace or a place of former education.
What it does
The bots identify people that would fall into this extremist category through the use of k-nearest values algorithm. One of these bots then deletes the extremist language, acting as a admin. The second bot converses with the person in question, attempting to identify if they are likely to lash out at their nearby neighbors. If this is recognized, the third bot sets up a website with new contacts for them to get in contact with and discuss things, thereby reducing the risk that said person will lash out at society.
How I built it
writing a machine-learning algorithm in python, alongside a webscraper, and then using ai.api to develop the second bot. We chose to develop it by using reddit as the basis for the data.
Challenges I ran into
Dealing with the fact that parts of the program have to be hosted in google's cloud, but other parts have to be run on the laptop, leading to a issue there. Also, a time-crunch near the end, as the project proved to be somewhat more ambitious and difficult than expected.
Accomplishments that I'm proud of
Finishing the program without too many cut corners.
What I learned
Google collab has seriously slow upload speeds. Things like having too new of an account can create unexpected logistical errors.
What's next for PosiTron