Similarity metrics

[csuksangium]

-Convert categorical data to numerical
-Computer euclidean distance between users
-Cluster users together based on similarities

*Note: Should we preemptively cluster together? or cluster upon request

[freddiev4]

x-ref #57

[freddiev4]

I'm not entirely sure if this is possible or even if clustering is the right approach. I think this was better off being framed as a classification problem

Also I don't think we have enough time for this