The wait for all data ninjas out there will end soon. In three days’ time, Venturesity will hit the grounds with its grooviest Big Data Hackathon on 1st February, 2014.
And without further ado we present – The Challenge!
Here’s what awaits you at the hackathon-
Use the Million Songs Dataset to discover, model and visualize interesting relationships between artist attributes and song audio metadata, alternatively build a backend for a web based service for personalization (ex. recommendation, playlist creation) using the various song and artist metadata.
To know more about Million songs dataset and its related issues, you can check this- Million Song Dataset.
So, you must be wondering what kind of projects you can take up in the event. We have a couple of ideas from our side to get your grey cells up and running, but don’t feel tied to these templates, let your imagination run wild.
1) Recommendation Service: Use the song meta data and the user taste data to evaluate the performance of different recommendation algorithms. Does item-item collaborative filtering give better than user-user? Are there hybrid algorithms that might give better performance? How can we use the song meta data along with the user taste data to make context aware playlist recommendations? Can we use the content attributes of the tracks to improve recommendation quality over purely collaborative methods. For more information/groundwork can check out The Preliminary Study on a recommender System for Million Songs dataset Challenge.
2) Automatic Tagging and Lyrics Analysis: We can also use the musicXmatch dataset, which gives lyrics for some of the songs in the main Million Songs Dataset. Combining user play counts for various songs, their metadata and lyrics is there a pattern to the song popularity? Are lyrics and sound attributes are a good predictor for the tags that a particular song would get? Can we implement a viable algorithm for music genre recognition?
You can go through Music Genre Classification for that!
3) Prediction – Song genre/tag popularity: Are there identifiable patterns in song popularity which depend on the year released and genre? Certain genres rise and fall during certain periods of time, can you make interesting visualizations which show these patterns? Can you predict a set of tags that would see a rise in popularity in the recent and coming years? Can you predict the outcome of the next Grammys? Can you predict the genre of a song by its audio data? You can go through the Final report of Music Genre Classification with the Million Song Dataset for more.
Having said that, there are a couple of other things we need to get out to you.
You do not need to qualify any criteria to be a part of this hackathon. Programmers, analysts, designers- all are welcome. Yes, you have heard it right- the event is free and there are no other charges involved! You can come as a team. However, if you are alone, it will not be an issue, we will find you a teammate!
This is Venturesity’s first Big Data Hackathon. Our in-house tech team and all of our instructors will be there to provide assistance and guidance for the participants. Various Big Data startups have been invited. So, you can look forward to getting connected while hacking. Awesome goodies and lot of media attention awaits the winners.
We are pleased to have on board Qubole as our infrastructure partner for the event. They are providing an elastic Hadoop cluster along with the full Million Songs Dataset on the cloud for our hackathon participants to work on.We also have Fever FM as our Radio Partner to cover our entire event.
Its exciting on our part to see more than 100 developers from product companies and startups register for the event. We are almost done with preparing the battleground for a hacker-night-out. Are you ready for it? Registration is still open. If you haven’t registered yet, the hurry!!
For any further help, drop an email to email@example.com
Date: 1st February,2014 – 2nd February, 2014.
Time: 10 A.M. – 10 A.M. (24 hours).
Venue Partner: Microsoft Accelerator,
Vigyan, #9, Lavelle Road,
Bangalore 560 001, India