Last week, I attended the first Hopper Elasticsearch Hackathon. This all day event was co-organized by Hopper, a Cambridge, Massachusetts, US based company that is using Elasticsearch to plow through a massive amount of travel data. Also organizing the hackathon were the fine folks from the Boston Elasticsearch Meetup, a diverse group of Elasticsearch users in the Greater Boston area.
The hackathon gathered more than 40 software engineers, students and other search and open source software enthusiasts eager to learn more about Elasticsearch and share their knowledge with others. The goal of the Hackathon organizers was to create an inclusive event that would be interesting to both experts and novices alike, so we started with short introduction into Elasticsearch to bring all attendees up to speed before getting started coding.
During a typical one-day hackathon, attendees have only 5-7 hours to write code, and getting stuck during these hours can derail otherwise interesting project. In order to improve participants’ experiences and allow them to make as much progress as possible on their projects in just a few short hours, the Hackathon organizers convened a group of local Elasticsearch experts from Hopper, Traacker and Elasticsearch Inc as mentors. I was one of the six mentors during this hackathon, and it was an amazing experience.
This event wasn’t my first Elasticsearch-themed hackathon, but what made it very different for me was the number of new Elasticsearch users and the amount of progress that they made during a very short period of time. This Hackthon was the first where we introduced the new client API and Kibana, and I think this two projects made all the difference. With the lightweight API, new users were able to very quickly translate concepts described in the elasticsearch.org user guide into working code. At the same time, Kibana made it simple to look at very complex data without requiring a lot of new code. In just seven hours by combining Twitter river with Kibana, one of the teams managed to create Twitter analytics solution for analyzing demographics of topics of interest (in their case tweets about Justin Bieber). Two winning teams analyzed Twitter to determine the popularity of TV shows and did sentiment analysis on political topics. But it wasn't all about Twitter; other hacks ranged from finding the shortest path between two Wikipedia articles to finding hidden relationships between Enron employees based on their email messages.
Two teams won prizes in this hackathon, and it was wonderful to be a part of something where everyone was learning stuff, having fun and left with a sense of accomplishment.
Many thanks to our sponsors: Elasticsearch Inc., SoftLayer and Traackr. You can read even more about the hackathon on the Hopper Blog.