found the key!!!

Analysis Excel Sheet

Ok finally some good news!!! Well it’s been ages now since I have been trying to tune last.fm to recognize my preferences, and it has disappointed me a lot. But off-late, I have seen remarkable improvement in the recommendations. Especially the artists who have popped up in the last few days have been phenomenally awesome!!! At one stage I had Saliva, Fort Minor, Mark Shinoda, Linkin Park, and some other awesome artists listed in my recommendation pool.

I think I can safely conclude now that there are 2 factors that drive the ‘goodness’ behind the last.fm recommendations. Firstly, the music you choose and rate should be compatible with current ‘popularity charts’, which are in turn derived from what other users have been listening to recently. Secondly, having the right neighborhood of people and groups of music lovers associated to your profile, who in turn influence how the popularity charts are over-layed on your preferences to make recommendations for you.

The songs that have been playing recently are very popular, because I have heard them on the radio a lot too. And I have rated a few of them to be good on my last.fm profile. This in turn dragged more artists and music of similar kind from the top of the charts into my recommendation list.

Recently a lot of my neighbors have been listening to artists who have been more or less close to my preference. Also the variations in their tastes have been few and far, keeping the threshold of variation in my recommendation pool to a minimum. This predictably makes my recommendations a lot more similar to what they have been in the past few weeks.

Running short of time here, so to conclude I would say that surrounding with good people and rating artists who matter most in the rating charts actually helps in fin-tuning my recommendations.

A final look at my analysis is necessary and I will do this in the coming few weeks. For now, I think I have found the secret key to unlock the potential of last.fm!!!