Preparing the Data Sets

I have started using the Echonest API to aggregate hotness ratings for a particular artist using the hotttnesss API call.

A useful API hook is similar. It basically takes an artist name as a seed value, optionally taking in min_hotttnesss and max_hotttnesss values. Then it returns all the artists similar to the one supplied. I plan to retrieve the hotness and familiarity ratings for the first 10 artists returned to prepare the first data set. Echonest's API documentation isn't much of a resource when it comes to explaining how hotness is actually calculated.¹

However, it is enough for me to work off of. By plotting the familiarity and hotness ratings I can get a correlation to begin with. Currently, I am have implemented Fruitr as a Rails app and started abstracting some Echonest API hooks. I am planning to deploy it using Capistrano, just as a learning exercise. Stay tuned.

Fruitr, last.fm and correlation

I have started a side project called Fruitr, which is an attempt to implement my idea of correlating information from two unrelated domains–music and population distribution of a city(tentative).

Currently, I am looking at ways to aggregate information by communicating with the Last.fm API¹ to create a music data set. I want to narrow down the domain to something like "Most loved tracks from June 2009 to September 2009" or "Most loved artist from June 2009 to September 2009." I also need to determine a sample size for this data set.

A simple example of what sort of data we can get using Last.fm:

def fruitify
    unless params[:name].blank?
      url = URI.parse("http://ws.audioscrobbler.com/2.0/?method=artist.getSimilar&artist=#{params[:name]}&format=json&api_key=#{API_KEY}")
      @result = Yajl::HttpStream.get(url)
    end
  render :layout => 'application' 
  end

Here I simply query the API to get a list of artists similar to the one I supply to it. The results are returned as JSON and I use the yajl-ruby gem² to parse that JSON into a ruby object.

This is the first in the series of posts which I will use to document the research and progress on the correlation project.

Year One

This post is intended to be a subtle indicator this site being born. Born out of a singular process that carried me through several sleepless nights. Born out of a stream of consciousness that refuses to settle upon anything less than a monument of perfection. Born out of that conflicting thought that all other websites are born out of the very same desire. Born in spite of that conflicting thought. Born to breathe life. Born to validate this post. Born to be wild (and other allusions to pop culture)

Born as a result of and to substantiate this quote by Marcel Duchamp:

The individual, man as a man, man as a brain, if you like, interests me more than what he makes, because I've noticed that most artists only repeat themselves.

I am done riding my high horse. His name is Horace. Horace shows up when it's time to race. He is a thoroughbred and will refuse to show up at any other time. It is in Horace's nature to allude to superlatives.

I have a tendency to think that Horace will also make for a fine war horse. Not that I would fancy myself fighting battles. However, this site, just like me, was born into strange times. Strange times mandate that I be prepared with the best of my defenses. This fact, combined with Horace's proclivity for racing and thinking very highly of himself, have all contributed to the birth of this site.

That's it for year one, horses, Horace in particular, websites and races.