Thanks, Moonbeam! Can't say I'm surprised by these showings, but really, there's a reason why these artists are acclaimed, right?

And here are the top 10 correlations again. The values have become closer to 1, since the agreement between voters is stronger in the top of the list.

1. Daniel & Stephan (r=0.62)
2. Daniel & Nicolas (0.57)
Stephan & Nicolas (0.57)
4. Nicolas & Honorio (0.50)
Harold Wexler & Rocky Raccoon (0.50)
6. Penguin & Slush (0.49)
7. Daniel & Rocky Raccoon (0.47)
8. Stephan & Miguel (0.45)
9. Stephan & Greg (0.44)
Jonah & Slush (0.44)
Harold Wexler & Kevin (0.44)

Looks like a group with a very similar taste!

However, I have also inluded Greg Rumpff in my stats, and then it looks like this...

1. Daniel & Stephan (r=0.62)
2. Daniel & Greg Rumpff (r=0.59)
3. Daniel & Nicolas (0.57)
Stephan & Nicolas (0.57)
Greg Rumpff & Nicolas (0.57)
6. Stephan & Greg Rumpff (0.53)
7. Nicolas & Honorio (0.50)
Harold Wexler & Rocky Raccoon (0.50)
9. Penguin & Slush (0.49)
10. Daniel & Rocky Raccoon (0.47)
Greg Rumpff & Greg (0.47)

Now, that's a really tight group! It would be easy to think that Daniel, Stephan & Co. are the most correlated with the overall list, but they're not. Here are the voters with most artists in the overall top 100.

1. Harold Wexler (73 artists)
2. Honorio (67)
3. BillAdama (64)
Kevin (64)
5. Henrik (61)
6. Greg (50)
7. Mitchell Stirling (48)
8. Fred (47)
Jonah (47)
Rocky Raccoon (47)

I've also made an attempt to show our tastes in a two-dimensional graph. I have used a method to find the two dimensions that best describe our different tastes (of course it's not possible to show all relationships between 36 voters with only two dimensions)

If you can't see the symbols proberly in the image below, you can click it to view it in a bigger size.

Henrik, I could not see the image, even after clicking on it. Perhaps this works:

I'm not sure if that graph made any sense to you. The question you would like to ask is probably "what are those two dimensions?" and to answer that I have - based on each voter's votes and position in the graph above - plotted the top 100 artists as well.

By looking at the images below the x-axis looks almost like a timeline, from Elvis Presley to LCD Soundsystem. The y-axis is a little harder, but I think that if you move upwards in the graph the artists become less and less guitar-based rock. However, this trend does not include jazz, as Miles Davis and John Coltrane are both in the lower region of the chart. An alternative interpretation is that the higher y-value the more the artists become dependent on other personnel, such as backing musicians and producers.

Perhaps you can come up with a better interpretation of the charts, but if I'm correct than the voters to the left are the biggest fans of "old music" while the the voters to the right prefer more recent music. The voters in the bottom (nothing wrong with that) prefer guitar bands while voters in the top are bigger fans of other genres/styles such as soul, rap, synth, etc.

The 10 artists with lowest values on the x-axis:
1. Elvis Presley
2. Chuck Berry
3. The Byrds
4. Van Morrison
5. Creedence Clearwater Revival
6. Neil Young
7. Johnny Cash
8. Hank Williams
9. The Rolling Stones
10. Elvis Costello

The 10 artists with highest values on the x-axis:
1. LCD Soundsystem
2. Massive Attack
3. The Strokes
4. Arcade Fire
5. New Order
6. Portishead
7. Blur
8. The Flaming Lips
9. Pulp
10. Joy Division

The 10 artists with lowest values on the y-axis:
1. R.E.M.
2. Pavement
3. The Clash
4. Talking Heads
5. Pearl Jam
6. Pixies
7. Wilco
8. Beck
9. Nirvana
10. Neutral Milk Hotel

The 10 artists with highest values on the y-axis:
1. Michael Jackson
2. Scott Walker
3. James Brown
4. Steely Dan
5. Prince
6. Kraftwerk
7. The Cure
8. Brian Eno
9. Super Furry Animals
10. Wu-Tang Clan






That's all for now...

Ooops, turn out very big, sorry about that.

I like this purple sun in the centre, that's me very nice. But what does this graph actually mean?

And oops again, there's already the answer to my question.

Andre, if you click on the image and a new window opens. There you need to click on the image again to maximize it.

(Or I've could have posted the images without resizing, just like you did, but they are too big for my screen...)

Henrik, it doesn't seem to work with Firefox. I just checked with IE, both image in the thread itself as well as popup works there...

My position in the graph matches perfectly, Henrik. I like both old and new music, but the peak lies in the 80's. And as you said, there is nothing wrong with guitar rock.

Thanks Andre!

I'd like to mention again that I'm really interested to hear if anyone has any other interpretation of the chart patterns.

Off to bed.

Hmm... Your interpretation seems pretty good to me, Henrik. It makes a lot of sense; I favor new music a lot more and am slightly on the guitar-based rock side. Cool.

Awesome graphs, Henrik! Could I ask in a little more detail how the x-axis and y-axis positions are determined? You say that the x-axis is a sort of timeline running from oldest to newest, based on artist. How are these artists ranked? Also, is the x-coordinate of each of the poster's positions based on a weighted average of his/her lists? That seems pretty cool.

I'm more unsure of how to interpret the y-axis. It seems to be based on genre, somewhat. How do you quantify something like that?

I'll send you the master list soon, Henrik. You don't have to, of course, but I think it would be really cool to see how it worked out if you included all of the artists rather than those who managed a top 100 place!

Moonbeam, I didn't go into any details of my method since it's a bit complicated, but I knew that you would ask me about it.

Out of all correlations, I took the most negative one (I didn't plan to reveal this, but what the heck, it was between you and Paul (r=-0.25)) and this was the start of the creation of three heterogenous groups; the Moonbeam group, the Paul group and the Inbetween group. The Inbetween group included the remaining 34 voters to begin with. Next step was to move 11 of them into the Moonbeam group and 11 others into the Paul group, so that each group had 12 voters. I moved one voter at a time to each of the groups, using the highest and lowest value of the difference between the 'mean correlation with the voters in the Moonbeam group' and the 'mean correlation with the voters in the Paul group'. When each group had 12 voters, the x-variable was complete; it's these difference values that are plotted in the graphs. And as you can see, in the end Moonbeam and Paul didn't have the most extreme x-values.

For the y-axis I went over this procedure almost in the same way, but I started by looking at negative correlations between voters within the Moonbeam group, within the Paul group and within the Inbetween group. Then I forced the three new y-axis groups to include 4 from each of the former x-axis groups.

I guess this is similar to factor analysis and there's probably a program that could have done this for me, but it was quite fun to do this manually...

What about the artist graphs then? Again I used the difference between the 'mean of the voters in the Moonbeam group' and the 'mean of the voters in the Paul group'. However, this time I used the mean position in our lists. For example, the mean position for Elvis Presley was much lower in the Paul group than in the Moonbeam group, which gave a negative value on the x-axis.

Uhmmm, did anyone understand anything?

I love it! "The Moonbeam Group"- awesome! Maybe I can sway a few more over to my side. Come to the Moonbeam Group.... Come....

I'm in the middle! Next to Sonofsam!

I've loved these graphs!
Now we have a map of all the member's distribution per musical taste. I totally agree with my position, and I'm actually next to people whose taste I consider a bit similar to mine.

As for the artists images, there are some interesting things to note. See that, while all the '60s acts are together (except VU, but that's quite obvious), The Beatles are in the alternative rock group. Maybe that occured because people with all different tastes like them, putting them in the exact middle of the list (and this would be the same case of U2).

Also, note that all the "white" music singer-songwrites are altogether, even though I see not so much similarities between, say, Costello and Dylan, for exemple.

I think that's me almost perfectly in the middle.

I was looking at the full 1,000 artists in order, and saw Curtis Mayfield somewhere in the 700s because it shows Vgrd spelling it Cutis, so Curtis Mayfield is really #97 instead of #101.

Georgie- you're right! Sorry, everyone, and Mr. Superfly!

#3 Radiohead. Does not appear in my list...at all. While I confess that repeated listens to OK COMPUTER and some of their albums have given a BIT more appreciation for Yorke and the lads..and I now actually like "Karma Police", "Fake Plastic Trees" and "House of Cards" (as well as "Creep" which I always liked) I still feel they're VASTLY overrated. Maybe IN RAINBOWS is the album that changes that for me...I'm liking it more than I have any other Radiohead so far.

#2 Dylan was my #11. Right now I only own compilations by him but a copy of BLOOD ON THE TRACKS and HIGHWAY 61 REVISITED (or is it HIGHWAY 51?) are in my future. It seems I like a LOT of his songs.

#1 The Beatles were my #2 just behind the Beach Boys. If you think that's sacrilege, ask Paul McCartney what the best album ever made was. Hint: He will NOT say it was a Beatles album.

The leap from OK Computer to Kid A is possibly the greatest change in the direction of a band, ever. Each of the guys in Radiohead basically had to re-learn their role within the band, which I can't imagine gives a warm-and-fuzzy feeling after you've just struck artistic gold with an album of such immense popularity and acclaim. If that's overrated... well, shit... where does that put all of the bands of the last 15 years who HAVEN'T released a seminal piece of dystopian pop art and then have done a complete one-eighty into apocalyptic electronica?

The Curtis Mayfield correction unfortunately means that LCD Soundsystem gets pushed out of the top 100. So it goes. "When someone great is gone..."

I'm so pleased that Curtis (belatedly) made it! Though it's too bad an important '00s artist like LCD gets dropped? Maybe we can just agree to asterisk Queen out of the list.