![\"\"](\"http:\/\/eee.uci.edu\/clients\/bjbecker\/ExploringtheCosmos\/almagestarab2.jpg\")
<\/p>\n","protected":false},"excerpt":{"rendered":"I need some math\/statistics help. I’m trying to figure out with some spreadsheet mojo whether math can give me an insight into who my favorite artist is. (I think I know the answer, but I’m open to being surprised by what statistics might reveal to me). If you’ve ever been interested in figuring out statistically who your favorite artist is, and you use last.fm (audioscrobbler) in addition to MOG , then read on. Mathemagicians welcome.<\/p>\n
MOG has given me a list of the 50 most popular artists\/bands in my collection according to the number of tracks I have for each of them. (There’s a widget that indicates this if you haven’t noticed) The “popularity” of this chart is skewed by artists who have many short tracks (such as J.S. Bach), or many tracks that I just don’t listen to that often (such as John Williams).<\/p>\n
Meanwhile, my charts at Last.fm\\audioscrobbler provides a list of my 50 most popular artists\/bands according to how often I listen to them. Since I listen to so much of my music using shuffle, the “popularity” of this chart is skewed again by those artists in my collection who have many tracks (i.e., J.S. Bach). My listening isn’t entirely random however, since I often skip albums by artists I’m not interested in listening to.<\/p>\n
If I can compare “artists with most tracks” with “most listened artists” I should be able to locate the outliers and get a true sense of who I really most like.<\/p>\n
I set up two columns with the 50 most listened to artists (via last.fm) in one column, and the 50 artists with the most tracks in my collection (via MOG ). After sorting the two columns and arranging the artists so that they appear alphabetically I have their two rankings side by side. (For example: J.S. Bach ranks 15th in most tracks I have, and 5th in most listened to. Other artists such as Legion of Green Men, are ranked in the popular listening column but don’t even make MOG ‘s cut in number of files).<\/p>\n
Now I need to perform some equation on the two popularity values to determine a most popular score. Any ideas?<\/p>\n
Assumptions: Artists that score high in listening but low in tracks should have their scores amplified.<\/p>\n
<\/p>\n","protected":false},"excerpt":{"rendered":"