Many Universities have offices like The Office of Career Consulting here at UPenn, yet these resources are often criminally under-maintained and underused. Meanwhile, students and job-seekers across the country are unsure about the industry standard - what should I do to get a career? What have others done to get the career I want? What is expected of me?
What it does
ProjectMe (theoretically would) operate off the proprietary data set of careers we mined using Selenium and Python during PennApps to return users relevant career data. Specifically, the system could be queried for paths from one career to another, for common paths from a career, and other interesting data visualizations. This could then be served to the user via interactive data interaction to best help the user understand what others in his/her field are doing to get employment. Furthermore, the system could connect to other users on the site (if they opt in) for light mentoring and networking opportunities.
How we built it
By generating a graph of all possible careers and one directional vertices from career to career (and sorting individual resume elements into careers using Tensorflow topic modelling) we can generate a dataset of paths between different given careers. Then we can walk this network with an algorithm not unlike Google map's distance algorithm, except optimizing for longest distance (most amount of people on edges walked). This produces the most popular path. We do this for a few paths and serve these paths to the user.