We were thinking about big data, and how it can be used to stop crime before it happens. There are systems currently in place to search for terrorism keywords, however we focused our on efforts on already offending people, and focus more on violent/malicious words.

What it does

Our program goes through a .CSV file of sexual offender names and details, and searches for them on twitter by name, and once it finds someone, it looks through their most recent tweets and flags tweets with concerning keywords.

How we built it

We used the Twitter API, with the python-twitter module, as well as the public data on sexual offenders to build our entire project on python.

Challenges we ran into

Confirming whether it is the right user or not, which we tried fixing by only counting tweets of names that had less than three accounts. Learning how to use the Twitter API, and also running into problems with the rate limit of the API. Also some accounts were set to private so we had to skip those.

Accomplishments that we're proud of

It is able to retrieve tweets and search for keywords, and we have already found a couple tweets that fits our danger criteria.

What we learned

We learned how to use the Twitter API, and how to parse data from a .CSV file.

What's next for Predictive Tweet Risk

We well expand the keywords, and make a website that has these flagged tweets embedded so that a human can make their own conclusion about the tweets and act accordingly.

Built With

Share this project: