D E Shaw hackathon which started in 2012 (& was a huge success), was conducted again this year. Like last year, i was very enthusiastic & jumped in to participate in a team of 4 people (including me). We named our team as baZZinga, which was our way of saying “zing!” like when you zing someone in a joke.
The theme of this year’s hackathon was out a week before the actual event would begin & as expected lot of teams started brain storming to come up with some innovative ideas. Unlike other teams, we didn’t discuss anything until the last day (the good old engineer in us never dies!). We met just 14 hours before the event was to begin & put forward all the little ideas that we had. A bunch of the ideas were good-to-have but not-very-innovative/useful. Our ideas ranged from developing a framework to display any standard data visually (charts/graphs) to developing an employee evaluation tool. We finally zeroed in on the employee evaluation tool as our project for this year’s hackathon, after which we disbursed.
Snapshot of what we discussed
Th event was to begin at 10 AM. We reached the venue well before time (for a change!) & waited for the countdown. The clock tick 10 & all the teams started off with the work. We chose a cubicle with 4 work stations & settled down with some yummy snacks. After the food & some videos on youtube, we decided on the important modules in our to-be-coded project & distributed the work equally, after which we worked seriously (well, with lot of fun in b/w) for sometime. In between we had a couple of mentors visiting us, with whom we shared our project idea, to which they gave a positive feedback (may be, this increased our confidence a bit). By lunch we had a web application skeleton ready & committed to our repository, after which we started work on the core modules. We did a lot of research & zeroed in on a NLP library & an open source sentiment analysis library. Before putting them to use, we tested them on some sample data & to our surprise the results were pretty good, hence we decided to use them for the real data. This was the time when we decided on the data sources to be used for evaluating an employee, we then extracted data from these data sources & integrated it with the NLP sentiment analysis libraries. The output had an accuracy of about 70%. By the time we did this, it was already 9. I felt very tired & decided to leave, promising my team members to come very early
the next morning, thus day 1 ended with some good amount of work being completed.
This was the THE day – we had to finish our work by 5 in the evening & present the work before the jury in a 2 minute presentation. I woke up late (hence, couldn’t keep up my promise ), but after reaching office we put in some serious work in adding UI for the data work that we did yesterday. Deciding the UI components has always been a challenge in S/W development world & not surprisingly it took good amount of time for us to decide on how to present the data. We decided on presenting the data using charts/bar graphs/line graphs. We re used Highcharts, one of the rich chart libraries to build our UI. We almost finished our work by the end & prepared for the presentation. Now was the time for presentation & we had some brilliant ideas by various teams, one of my team members presented the key aspects of our project to the jury. This ended our HACKATHON journey.
The results were announced a week after the presentations. Frankly speaking, we weren’t expecting anything, but the results are always unexpected , we won the Hackathon for best creative idea. To add cherry on the cake, we were given the prize (XBOX) by well known actor/director Rahul Bose.