One of my favourite things about my Last.fm subscription is it’s tracking of my listening history. It’s one thing to try and arbitrarily choose today what my top songs of 2009 were, but to actually look at what tracks I played the most over the year offers a more interesting (to me anyway) view.

So here’s the list of my top 25 songs of 2009, in order of the number of times I played the full track on my computer, iPod, or iPhone:

1. Joel Plaskett Emergency – Nowhere With You
2. Flogging Molly – What’s Left of the Flag
3. Hi-Standard – Wait for the Sun
4. Portal – Still Alive
5. Leonard Cohen – Waiting For The Miracle
6. Enter the Haggis – One Last Drink
7. Great Big Sea – The Night Pat Murphy Died
8. Jonathan Coulton – Mr. Fancy Pants
9. The Salads – Get Loose
10. Styrofoam and Sarah Shannon – I Found Love
11. Jonathan Coulton – The Future Soon
12. Finnegan’s Lads – Dirty Old Town
13. Da Vinci’s Notebook – The Gates
14. Semisonic – Closing Time
15. Jonathan Coulton – Code Monkey
16. Three Dead Trolls in a Baggie – The System Administrator Song
17. Dean Elliott & His Big Band – Lonesome Road
18. Corb Lund Band – The Truck Got Stuck
19. The Tossers – Altercations
20. John Coltrane – Giant Steps
21. Dropkick Murphys – The Dirty Glass
22. Patti Smith – Gloria
23. Tenacious D – The Metal
24. Rancid – Roots Radicals
25. Flogging Molly – Rebels of the Sacred Heart

Of course, the math geek in me just screams at the idea of using raw play count data that doesn’t take into account when any of these songs actually became available to me. For example, I bought Enter the Haggis in November, and they made it to #6. I bought Jonathan Coulton’s album in March, and he appears three times (four if you count “Still Alive”). And yet, bands like Flogging Molly and Corb Lund, oft-stated “favourite” bands of mine of which I own many albums, only appear once or twice. Some of these tracks I don’t own at all, but have appeared often enough as “recommendations” on my last.fm station that they make the list!

Hmm… I wonder how hard it would be to build an algorithm that could level out the initial enthusiastic repetition of a new album or song, consider the time of year it was added to the overall library, and assign an appropriate weighting as compared to those that have been in rotation for more than 12 (24? more?) months. I wonder how clear the pattern would be if I could go back and track plays of a new song from purchase, peak, trough, to plateau. What kind of timeline does there need to be for a song to stabilize into regular rotation, or just be forgotten until it gets picked up by a random shuffle and brought back into play? How much of an effect does placing the song into a playlist have versus leaving it in the overall library?

How much time can I waste discussing pattern analysis of my music listening habits before I get back to work?

That one I think I can answer.