Page 1 of 2

Artist Rating Correlations

Posted: Sun Jul 17, 2016 8:11 am
by Moonbeam
A few of us have been rating the artists unveiled in the all time artist poll, and I thought it would be fun to look at correlations between users. With that in mind, I have put together this Google Sheet with 325 artists for us to rate to our hearts' content.

These 325 artists all feature in at least one top 200 out of the 4 AM artist polls and the AM artist rating. If you'd like to participate, add your name to a new column in Row 1 and then rate as many artists as you'd like. You can use whatever scale you like: I've gone for 0-10 ratings, but whatever numerical scale you'd like should be ok, as long as the assigned numbers are in a natural order. Whether you care to rate all 325 or just a subset, I'm happy to calculate correlations (both the standard Pearson and a non-parameteric) between users and post the results.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 8:41 am
by Romain
Check !

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 9:43 am
by Bang Jan
Added mine.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 11:31 am
by EmilienDelRey
Added mine too.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 11:32 am
by Moonbeam
Cool to see the sheet grow! I added a sheet that manually calculates the Pearson correlations (well, for the first 20 user columns - I got tired after that).

It seems to me we could come up with an automated recommendation system if we did something similar for albums and songs. Perhaps once the site updates, we could set up similar sheets with ratings, and each user can use, say, the strongest 5 user matches to "predict" ratings for unheard albums/songs.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 12:43 pm
by jamieW
I've added mine, as well. This site can sure be humbling at times! Going in, I assumed I'd be at least passingly familiar with all of the artists, but there were a few I don't remember ever hearing anything from. There were also some others that I have heard but didn't rate, since I felt that I wasn't familiar with enough of their music to judge.

Thanks, Moonbeam!

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 1:13 pm
by Duncan Thaw
That was good fun - a game which works for those of us with mere-mortal levels of music knowledge.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 4:57 pm
by Bruno
Added mine.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 5:15 pm
by Nick
Browsing the list I've seen a couple 0's for some artists, which really makes me wonder, what exactly would constitute a 0?

I don't think there's any artist I dislike enough to give a 0 to. My least favorite musical act I've ever heard would probably be Brokencyde (if you haven't heard them, check them out, you're in for a real treat), and even then I'd probably give them a 0.5 or something. But a pure 0? If you claim that The Rolling Stones (or whoever) are a 0, then what does that say about something like neo-Nazi punk rock? Is that on the same level of "bad" to you as The Rolling Stones? Do you start giving out negative numbers?

Giving out 0's also seems to go against my main philosophy that I use to rank artists, where I only really count "the good stuff" when I factor in an artist's ranking. Therefore, I feel confident giving a positive rating to an artist I love as I assume I've heard just about everything there is to love about them, and any bad music they've put out that I haven't heard is irrelevant. But to give an artist a 0 suggests that you've explored their entire discography and found not a single thing to raise their score higher than nothing.

I guess I just don't understand how someone could listen to a massive amount of music from any given artist and decide there is absolutely nothing of any redeemable value to be found. I could even say a couple good things about Brokencyde if need be. As absolutely awful as their music is, they do have a somewhat unique sound to them, and the fact that it doesn't seem like they're taking themselves too seriously is at least somewhat endearing. Don't get me wrong though, this is about as bad as music gets, (baring music that's blatantly racist, sexist, homophobic, etc.) but even then I can still find some good in it.

How do you guys feel about this?

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 5:30 pm
by cjppats43
I'm going to put my answers in now

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 5:31 pm
by jamieW
Nick wrote:Browsing the list I've seen a couple 0's for some artists, which really makes me wonder, what exactly would constitute a 0?

I don't think there's any artist I dislike enough to give a 0 to. My least favorite musical act I've ever heard would probably be Brokencyde (if you haven't heard them, check them out, you're in for a real treat), and even then I'd probably give them a 0.5 or something. But a pure 0? If you claim that The Rolling Stones (or whoever) are a 0, then what does that say about something like neo-Nazi punk rock? Is that on the same level of "bad" to you as The Rolling Stones? Do you start giving out negative numbers?

Giving out 0's also seems to go against my main philosophy that I use to rank artists, where I only really count "the good stuff" when I factor in an artist's ranking. Therefore, I feel confident giving a positive rating to an artist I love as I assume I've heard just about everything there is to love about them, and any bad music they've put out that I haven't heard is irrelevant. But to give an artist a 0 suggests that you've explored their entire discography and found not a single thing to raise their score higher than nothing.

I guess I just don't understand how someone could listen to a massive amount of music from any given artist and decide there is absolutely nothing of any redeemable value to be found. I could even say a couple good things about Brokencyde if need be. As absolutely awful as their music is, they do have a somewhat unique sound to them, and the fact that it doesn't seem like they're taking themselves too seriously is at least somewhat endearing. Don't get me wrong though, this is about as bad as music gets, (baring music that's blatantly racist, sexist, homophobic, etc.) but even then I can still find some good in it.

How do you guys feel about this?
Obviously, it's up to each individual to determine how they choose to rate these artists, but, personally, I agree with this. (Remember, my description of my personal rating scale had Michael Bolton, my all-time worst artist as a 1.) In all my many, many years of listening to music, I've only given one song/album/artist a 0 and that was a song from the '60s called "An Open Letter To My Teenage Son" by a Grand Rapids, Michigan DJ. (It actually hit the Top 10 on the Billboard chart; and, of course, he would have to be from my home area. At least, I have the pleasure of telling people Sufjan Stevens attended college in my hometown, which balances this out somewhat.) It's been a long time since I've heard the "song," which is really a spoken-word record, but basically it's a dad telling his son that if he doesn't go fight in the Vietnam war, he disowns him. I found it so utterly amoral and vile that it's the only piece of music I've ever given a 0. It's why I give even my least favorite songs/artists/albums a 1. For me, there needs to be a way to separate awful music from morally reprehensible music, but I can understand how others may see it differently.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 5:34 pm
by prosecutorgodot
Nick wrote:Browsing the list I've seen a couple 0's for some artists, which really makes me wonder, what exactly would constitute a 0?

I don't think there's any artist I dislike enough to give a 0 to.
I think there is absolutely no rational and logical way to defend labeling an artist a 0. You're basically saying "this artist has provided no one with happiness or pleasure, not one ounce of listenable music."
EDIT: Okay, I see one possible way to rate an artist a 0, and that is to commit suicide after listening to the artist's best song, in which case, you would be in no shape to type online a 0. I apologize if this is too harsh, but rating an artist a 0 is harsh and nonsensical!
Me personally, I don't see myself ever rating an artist anything below a 3, because that would mean that I listened thoroughly to their discography, and vomited and slit my wrists many times as a result, before I came to a "2.5". Thus far, the lowest I have rated is 4.0

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 8:30 pm
by Jirin
Do I need to move the averages column to the next column or will that happen automatically when I add ratings?

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:11 pm
by Moonbeam
As far as rating an artist 0 goes, it's just a personal scale. There is no way to ensure that everyone's rankings mean the same thing anyway.

For me, a rating of 0, just as a rating of 10, does not have any meaning outside of my own personal enjoyment. I'd never say an artist I'd rank a 10 is someone "everyone must hear" or "has something for everyone", just as I wouldn't say an artist I rank a 0 "has no value to anyone". All of my ratings only apply to me.

An artist to whom I give a 0 is one where every song I've heard is off-putting. I understand it's possible that such artists could have some songs within their discographies that I could tolerate, but that's why ratings such as these are always subject to change. 10 years ago, I probably would have rated Tom Waits around a 5 and Arcade Fire around a 9.5, but my feelings have changed since then.

I didn't intend to offend anyone with ratings of 0. If someone really likes Beyoncé or AC/DC, that's their right! The great thing about art is that there is someone in the audience deeply moved by it. Consequently, all art has inherent value.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:13 pm
by Moonbeam
Jirin wrote:Do I need to move the averages column to the next column or will that happen automatically when I add ratings?
I'm not sure what you mean by an averages column. Do you mean the correlations tab? That should update automatically.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:26 pm
by Neil
Please help!! I've added my column, but no ratings yet. Am I too rate all artists or just my top 100?

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:30 pm
by jamieW
Moonbeam wrote:As far as rating an artist 0 goes, it's just a personal scale. There is no way to ensure that everyone's rankings mean the same thing anyway.

For me, a rating of 0, just as a rating of 10, does not have any meaning outside of my own personal enjoyment. I'd never say an artist I'd rank a 10 is someone "everyone must hear" or "has something for everyone", just as I wouldn't say an artist I rank a 0 "has no value to anyone". All of my ratings only apply to me.

An artist to whom I give a 0 is one where every song I've heard is off-putting. I understand it's possible that such artists could have some songs within their discographies that I could tolerate, but that's why ratings such as these are always subject to change. 10 years ago, I probably would have rated Tom Waits around a 5 and Arcade Fire around a 9.5, but my feelings have changed since then.

I didn't intend to offend anyone with ratings of 0. If someone really likes Beyoncé or AC/DC, that's their right! The great thing about art is that there is someone in the audience deeply moved by it. Consequently, all art has inherent value.
Moonbeam, I didn't mean to imply there was anything wrong with you rating artists as a 0 on an individual scale. When I spoke, I was just agreeing with Nick that, for me personally, I cannot rate an artist/song/album a 0 unless there's something morally offensive to me. You're rating these on an enjoyment scale, which is entirely different. (And you certainly haven't offended me in the slightest.) You've gone out of your way, probably more than anyone else, to attempt to listen to and understand the acclaim of the artists you don't care for, and that's more than anyone could ever ask.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:30 pm
by Moonbeam
Neil wrote:Please help!! I've added my column, but no ratings yet. Am I too rate all artists or just my top 100?
Rate whichever artists you like - no rules! The more you rate, though, the more meaningful the resulting correlations will be.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:45 pm
by Neil
Moonbeam wrote:
Neil wrote:Please help!! I've added my column, but no ratings yet. Am I too rate all artists or just my top 100?
Rate whichever artists you like - no rules! The more you rate, though, the more meaningful the resulting correlations will be.
Thank you Moonbeam. Is there a standard rating scale? Or just a scale of my choosing? Sorry, I'm so needy!! ha ha

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:46 pm
by Moonbeam
Neil wrote:
Moonbeam wrote:
Neil wrote:Please help!! I've added my column, but no ratings yet. Am I too rate all artists or just my top 100?
Rate whichever artists you like - no rules! The more you rate, though, the more meaningful the resulting correlations will be.
Thank you Moonbeam. Is there a standard rating scale? Or just a scale of my choosing? Sorry, I'm so needy!! ha ha
Any scale you like! Most seem to be using a maximum of 10, but there is no need to do this.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:52 pm
by Jirin
Moonbeam wrote:
Jirin wrote:Do I need to move the averages column to the next column or will that happen automatically when I add ratings?
I'm not sure what you mean by an averages column. Do you mean the correlations tab? That should update automatically.
When I started my ratings, the column immediately to the right of the other user's columns was set up to take a mean of the other user ratings.

Re: Artist Rating Correlations

Posted: Sun Jul 17, 2016 10:55 pm
by Moonbeam
Jirin wrote:
Moonbeam wrote:
Jirin wrote:Do I need to move the averages column to the next column or will that happen automatically when I add ratings?
I'm not sure what you mean by an averages column. Do you mean the correlations tab? That should update automatically.
When I started my ratings, the column immediately to the right of the other user's columns was set up to take a mean of the other user ratings.
That's strange! Maybe someone was curious and added a column of averages? This would be a cool feature if everyone adhered to a common scale. How do people feel about me adding such a column to the left of the user votes?

Re: Artist Rating Correlations

Posted: Mon Jul 18, 2016 12:32 am
by Moonbeam
15 sets of votes in, and so far, the strongest (positive) correlations for each person:

Moonbeam: Romain (0.420)
Romain: Moonbeam (0.420)
Bang Jan: sonofsamiam (0.790)
Emilien: BleuPanda (0.639)
jamieW: Bruno (0.485)
Duncan Thaw: cjppats43 (0.604)
Bruno: jamieW (0.485)
luney6: Bang Jan (0.478)
JWinton: Duncan Thaw (0.571)
prosecutorgodot: Henry (0.294)
Henry: Bruno (0.463)
cjppats43: Duncan Thaw (0.604)
BleuPanda: Emilien (0.639)
Jirin: JWinton (0.527)
sonofsamiam: Bang Jan (0.790)

Re: Artist Rating Correlations

Posted: Mon Jul 18, 2016 9:47 am
by Moonbeam
I see that someone tried to add an averages column to the right of the rating columns. I've added average, variance, and # of votes columns as columns G, H, and I to the left for convenience.

Re: Artist Rating Correlations

Posted: Mon Jul 18, 2016 10:29 am
by spiritualized
Mine's done :) thanks for this, Moonbeam. It's like RYM for the forumers :)

Re: Artist Rating Correlations

Posted: Mon Jul 18, 2016 11:39 am
by Nassim
I expected it to be spiritualized, but looks like BleuPanda is my best Buddy (though to be fair spiritualized rated much more artists so it might be more relevant with him).
And prosecutorgodot is my nemesis (though here too, given the number of artists rated, my negative correlation with Henry might make more sense, and is not a big surprise)

Re: Artist Rating Correlations

Posted: Mon Jul 18, 2016 12:08 pm
by BleuPanda
I stopped after a bit because someone reorganized the sheet while I was working. I'll likely add more later.

Re: Artist Rating Correlations

Posted: Mon Jul 18, 2016 12:45 pm
by spiritualized
Nassim wrote:I expected it to be spiritualized, but looks like BleuPanda is my best Buddy (though to be fair spiritualized rated much more artists so it might be more relevant with him).
And prosecutorgodot is my nemesis (though here too, given the number of artists rated, my negative correlation with Henry might make more sense, and is not a big surprise)

Hahaha :) Yes I had a feeling we had a lot of common musical tastes. Although, bizarrely, you come up as a perfect correlation on mine, a 1.000 ! (computer says no...)
Prosecutor is also my nemesis :)

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 5:02 am
by whuntva
Locked mine in

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 5:23 am
by Moonbeam
We're now up to 22 sets of ratings! Neil, your column looks lonely there with 0 ratings. Care to fill it in?

For those who have participated, would you also be interested in similar ratings for albums and songs? I think it could be a cool way to generate recommendations. Basically, we could set up similar spreadsheets with the albums and songs listed at AM, and have another sheet with other albums and songs that users fill in with ratings for the purposes of recommendations.

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 11:56 am
by jamieW
Moonbeam wrote:We're now up to 22 sets of ratings! Neil, your column looks lonely there with 0 ratings. Care to fill it in?

For those who have participated, would you also be interested in similar ratings for albums and songs? I think it could be a cool way to generate recommendations. Basically, we could set up similar spreadsheets with the albums and songs listed at AM, and have another sheet with other albums and songs that users fill in with ratings for the purposes of recommendations.
I would definitely participate, but it might be a while before I could fill out the column for albums, since I'm in an eternal process of trying to rate all the albums I've heard throughout my lifetime. (Not to mention, I'm supposed to be hosting the early 20th century music poll right now. But then that's one of the many great things about the Google spreadsheet idea - people can fill it out whenever they have time.) I'm moving this week because of my job, but I could work on the song ratings once I get settled in. Thanks, Moonbeam! This is an outstanding project, and will be a tremendous source for recommendations.

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 12:24 pm
by BleuPanda
I would totally be down for albums or songs; I have a lot more solid of an opinion on specific works than I do on artists as a whole.

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 12:58 pm
by Moonbeam
Cool! I think I'll put something together for albums and songs once the new updates come through. :happy-partydance:

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 1:01 pm
by BleuPanda
Since I don't use google sheets, are there safety features here? Like, what's stopping someone from erasing everything? (I'm assuming there must at least be backups?)

Re: Artist Rating Correlations

Posted: Wed Jul 20, 2016 1:11 pm
by Nassim
BleuPanda wrote:Since I don't use google sheets, are there safety features here? Like, what's stopping someone from erasing everything? (I'm assuming there must at least be backups?)
Google makes regular saves so if everything disappears, you can restore an older version.
However if I were to make minor changes, say change Moonbeam's rating of the Beatles from 1 to 10, nothing can stop me and it might not be noticed right away :D

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 6:17 am
by Henry
Suggested corrections:

Clovers should have AM Rating of 1258.
Counting Crows should have AM Rating of 1013.

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 9:31 am
by Henrik
Henry wrote:Suggested corrections:

Clovers should have AM Rating of 1258.
Counting Crows should have AM Rating of 1013.
I read this first thinking that it was the report error thread and couldn't understand what you were talking about! :D

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 10:02 am
by Moonbeam
Henry wrote:Suggested corrections:

Clovers should have AM Rating of 1258.
Counting Crows should have AM Rating of 1013.
I've added those in, along with the few others with AM rankings below 1000. Any other mistakes in the sheet (I'm sure there are some) can be edited by anyone who notices them. And no, Nassim, my rating of The Beatles does not count. :P

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 12:59 pm
by Dan
Just noticed on the sheet that I have a variance of 0.84, which makes me the only one (so far) with a variance of less than 1. I had to Google what "variance" means in statistical terms. So for relative stats dummies like me...

I gave nearly all of the artists on Moonbeam's sheet a score out of 10. In statistical language, I have a "mean" score (which is a central tendency, very similar to an average score - in this case, my mean/average score for all of the artists I gave a score to is around 7.5 out of 10). My understanding is that "variance" calculates the tendency to deviate from the mean. So if my mean is 7.5 out of 10 and my variance is 0.84, it means that I didn't tend to give a score much higher than 7.5 or much lower than 7.5. You can almost use the variance of 0.84 and either + or - it to the mean score (so I didn't often go higher than a score of 7.5+0.84 or lower than a score of 7.5-0.84).

[imgsize 200x233]http://s2.quickmeme.com/img/87/877be8bc ... 485e8b.jpg[/imgsize]

Don't call me "girl". But thank you, Ryan Gosling.

I guess what I'm trying to say is that I couldn't justify giving many of the artists a really low score (in fact, I only gave Linkin Park and X a score of less than 5 out of 10). I feel that even the artists on the list that I don't care for much have contributed more to my appreaciation of music than most of the shite you hear on commercial radio nowadays.

Earlier in this thread, people talked about giving a score of 0 to artists, songs or albums. The only "song" I've heard that I would give a score of 0 to is 4′33″ by John Cage. It's four minutes thirty-three seconds of silence, and it's intended to make the listener think about the sounds of the enviroment they are in, and about how any sounds may constitute music. It's even on AM as the #2409 most acclaimed song. Hmmm... call it an art concept, maybe, but not music. It's bullshit music. It has zero music or lyrics in it, so it gets a score of zero.

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 1:07 pm
by BleuPanda
See, with my logic, 4'33 is a 5/10 song, as a song deserving of a score lower than 5 would be time better spent listening to nothing at all.

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 2:36 pm
by Henrik
Then, with my logic, only 2408 songs are worth listening to.

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 2:50 pm
by BleuPanda
Henrik wrote:Then, with my logic, only 2408 songs are worth listening to.

Exactly! :D

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 3:30 pm
by Henry
Looks like Bowie is the highest rated artist in this spreadsheet so far.

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 3:59 pm
by BleuPanda
Henry wrote:Looks like Bowie is the highest rated artist in this spreadsheet so far.

Our current top 25 based on average score:
1. David Bowie
2. Radiohead
3. The Beatles
4. The Rolling Stones
5. Stevie Wonder
6. Jimi Hendrix
7. The Cure
8. The Velvet Underground
9. Miles Davis
10. Prince
11. Led Zeppelin
12. The Kinks
13. Bob Dylan
14. The Smiths
15. Nirvana
16. The Beach Boys
17. The Clash
18. James Brown
19. Neil Young
20. Sufjan Stevens
21. John Coltrane
22. Aretha Franklin
23. Pink Floyd
24. Bruce Springsteen
25. Marvin Gaye

It's interesting how much our group opinion can change once we begin to both incorporate negative opinions and remove the influence of those over bands they haven't listened to. However, unless there is some sort of weight system, there's still a noticeable flaw created by some of us going from 0-10 while others go from 5-10 (those with a larger scale used would have a heavier say on the average).

Re: Artist Rating Correlations

Posted: Thu Jul 21, 2016 4:01 pm
by spiritualized
And Linkin Park has the lowest mark (amongst the ones that have significant votes, ie, more than five)... a paltry 3.9 / 10 !

Surprised at one thing : Stereolab (one of my favourite bands ever) getting half of the total votes. A band worth visiting for our voters ?

Re: Artist Rating Correlations

Posted: Sat Jul 23, 2016 11:51 pm
by Zombeels
Rating artists is very hard. Case in point. Fleetwood Mac. Love the early Peter Green era, the middle Weston & Welch years are in the middle and save for a couple of Lindsey Buckingham tunes the later years are dreadful.

Re: Artist Rating Correlations

Posted: Sun Jul 24, 2016 1:07 am
by Zombeels
spiritualized wrote:And Linkin Park has the lowest mark (amongst the ones that have significant votes, ie, more than five)... a paltry 3.9 / 10 !

Surprised at one thing : Stereolab (one of my favourite bands ever) getting half of the total votes. A band worth visiting for our voters ?
I actually own about 10 stereolab CDs but have yet to listen to most of them.

Re: Artist Rating Correlations

Posted: Sun Jul 24, 2016 3:31 pm
by Jirin
What amazes me is that some people have correlations up near .8. I've got in the .5 range with JWinton and Bang Jan, but a correlation of .790 means almost two thirds of the entire variation in your ratings can be explained by each others' votes. Seems extremely unlikely that any two people's votes would be that closely related.

I'd like to see multiple regression done on this data set.

Re: Artist Rating Correlations

Posted: Sun Jul 24, 2016 4:51 pm
by Henry
I am looking for song recommendations for the following artists on the list:

1) Curt Boettcher (rated highly by BangJan and Sonofsamiam),
2) George Jones (rated reasonably highly by JamieW and Jirin)
3) Jellyfish (rated highly by sonofsamiam and spiritualized)
4) Low (rated highly by Nassim and spiritualized)
______________________________________________________________________________________________

Perhaps Moonbeam can explain the criteria used to include the following artists on the list:
1) Geoff Moore
2) Handsome Family
3) High Llamas
4) Mojave 3
5) Why?

Re: Artist Rating Correlations

Posted: Mon Jul 25, 2016 9:54 am
by Nassim
Henry wrote:I am looking for song recommendations for the following artists on the list:
4) Low (rated highly by Nassim and spiritualized)
There is almost no wrong starting point with Low (where, there are a couple I guess).
I agree with critics and consider Things We Lost In The Fire their best album, it's as intimate and beautiful as anything they released before but also has some of the warmth and instrumentation of their latest albums.
That being said, I Could Live In Hope, their first album, might be the most influential one ; but it can also seem too sparse and slow, that's pretty much as minimalist as rock gets which could be boring to some.
"Knowing" your tastes, I might suggest to start with either C'mon which keeps their signature sound while being a bit more sonically diverse (I guess some would suggest the Great Destroyer for the same reason).

If you're not aiming for albums but songs, here are 10 favorite (chronological order)
Words
Missouri
Sunflower
Dinosaur Act
July
Point of Disgust
Witches
Especially Me
Holy Ghost
No Comprende