At a cybersecurity competition last week, it became apparent how much sensitive data people accidentally share on the web. This information can be used by bad actors to compromise your security by performing password resets.
What it does
This application stands your Facebook profile for personal identifying information or other sensitive details. The program first goes through all the users statuses and ensures that the user is not sharing any information that could lead to such as vulnerability.
We also generate other statistics including a brief summary of all your posts.
How I built it
We used Azure to run the webpage and Python flask as the back end. We used natural language process toolkits and HP's entity extraction API.
Challenges I ran into
Debugging 500 errors. Using CI on Azure.
Accomplishments that I'm proud of
We got the natural language processing working including generated summaries.
What I learned
How to do some simple language processing, how to build a flask site, how to load a flask site on Azure, how to configure Azure, how to use HP's Haven-Now API, and how to setup a reliable AJAX API.
What's next for webTLDR
Using Bayesian inference, logistic regression and other probabilistic methods to predict what you would say in a post.
Clean up the interface, make it look pretty.