andrewyli/hackintoit2016

ConnectMe

Created for Hack Into It @ Berkeley, Fall 2016

This is a prototype app to take in a user's LinkedIn and Facebook profiles, and generate a list of friends that can be used to ask for referrals to a job position within a company.

Contributors

Andrew Li Emma Hsu Jackson Stogel Jordan Hart Ken Luy

Project Goals and Model

Using linear regression on publicly available datasets such as that of the US Census, we created a model with the capacity to predict the stability and growth of an industry for a entry-level user of the application. Lasso was used in this process. Separately, use Facebook and LinkedIn API/CSV download/web crawlers to access friendlists and connections to get workplaces, and rate these (friend, job) pairs based on market viability. Return a sorted table of friends to get in contact with about finding a job based on a composite score.

Further Work

• Figuring out a much faster and surefire way to bypass the restrictions on getting a user's friends/connections, given that the two services of Facebook and LinkedIn fail to provide adequate APIs
• Potentially using more complicated modeling such as random forest regression in the future
• Given a faster and more reliable search method, provide true networking reach and allow access to friends of friends for cross-referral