Click your heels together and say the magic words three times: "there's no place like Yelp...there's no place like Yelp...there's no place like Yelp." Then perhaps you'll be whisked away to the land of Yelp Engineering where our third Hackathon just wrapped up. Like the previous Hackathon and the one before that, this 48-hour period where engineers were set loose to work on "anything you wouldn't normally be able to work on that might be useful, funny, or cool" (and ostensibly, related at least indirectly to Yelp) resulted in some awesome creations.
One of the biggest milestones for our third Hackathon was the inclusion of the first non-engineer participants! For instance, Alex O. from our Sales Ops team joined forces with engineer Aaron V. from our Systems team and A/V tech Patrick D. to produce an unofficial Yelp song:
Of course, there were plenty of software-oriented projects as well. One team created what they dubbed "MRPig" - inspired by the concept of Apache Pig, they created a general data query language on top of Yelp's open source MRJob framework that allows easier creation of MapReduce jobs for data analysis, with less need to involve an engineer in the process. Our product managers were excited about the possibility of having faster turnaround times on questions like, "is there a correlation between ad category determination accuracy and overall click-through rates?"