I noticed, on the following page (shortlistedjobs, one of the sites listed in your portfolio), that the "Distiller" takes the employee data and computes a list of factors that are then categorized into success and failure buckets. I see that being an introvert is considered a failure with a significance of -5. How do you quantify introversion? What method do you use? Something similar to sentiment analysis? I'm just curious.
http://www.shortlistedjobs.com/how-it-works/
Ha, this was one of my prior "failures" but being an introvert is not a "failure" per se. Shortlisted was designed to consume any information an organization had about an employee plus a 1 to 10 rating about the value of each employee to the organization.
The algorithm would then perform a step-wise regression on all the data and "distil" it down to a model of attributes that correlate with performance to that specific business, not in general. So being an introvert might be great for some business cases but may not correlate well with success in every case.
After the model was built, the second step in Shortlisted is to take another huge chunk of data and look for new employees that had attributes that correlate well with success for the business.
So a typical use case might be your company hires 10 engineers out of college per year. Out of those ten you can rank them in approximately highest to lowest in terms of value to the company. Shortlisted takes everything you know about those 10 (plus any from previous years, it really takes about 30 data points to produce meaningful results) and searches the University for more people like the most valuable employees and tries to avoid people like the least valuable.
Our goal was to help job seekers avoid situations that not going to be a good fit, and help employers find new hires that had a high likely hood of success.
