New mathematical paper shows songs spread like viruses

Originally published at: New mathematical paper shows songs spread like viruses | Boing Boing

2 Likes

Help, the Disturbed virus has spread to the Simon and Garfunkel host:

We have to keep the Scarborough Fair from coming Down with the Sickness :slight_smile:

2 Likes
4 Likes

I love data analysis stories and do a bit myself for fun and profit. The data set is fairly small and Electronica is just a niche in the genre’s in the study. I wondered what songs might be included in that group to see why they might have spread so quickly. I looked at the data in this file that seems to include the genre classifications. However it seems to be misaligned as the songs do not match the genre which leads to some quite amusing assignments.
This file seems to be correct.

count title artist genre r R0
5061 Take Me to the Hospital The Prodigy Electronica 646.1189424 141744.3002
4710 Stand Up The Prodigy Electronica 644.8788343 130937.2967
4700 Run With the Wolves The Prodigy Electronica 619.2408813 123290.6619
5534 Invaders Must Die The Prodigy Electronica 560.5049783 119480.1682
4817 Thunder The Prodigy Electronica 26.52798807 7562.479021
4286 Omen Reprise The Prodigy Electronica 19.49133873 3430.005788
5786 Omen The Prodigy Electronica 0.395112485 98.31938955
4669 Colours The Prodigy Electronica 0.502516833 8.900856539
9470 Starstrukk [Featuringkatyperry] (Explicit Bonus Version) 3OH!3 Electronica 0.144947976 7.332931196
4663 World’s On Fire The Prodigy Electronica 0.281246336 4.464098896
4636 Piranha The Prodigy Electronica 0.262159705 3.933227155
6419 Warrior’s Dance The Prodigy Electronica -0.03791741 2.94E-05

Out of the top 1000 songs that were analyzed in the Electronica section we are really talking about the singles release from the “Invaders Must Die” album by The Prodigy.
Personally I would classify Starstrukk feat. Katy Perry as pop rather than Electronica and it had a pretty low r number.

However these are not the songs that spread fastest, the top 10 sorted by R are here:

count title artist genre r R0
5273 You’re Never Over Eminem Rap/Hip Hop 2.21E+36
5039 What Is Love Take That Pop 2.33E+20
19694 Haven’t Met You Yet Michael Bublé Pop 4.21076E+18
5720 You Give Love A Bad Name Bon Jovi Rock 5962956568
5971 Said It All Take That Pop 277136258
5356 The Captain Biffy Clyro Rock 21745636.89
4752 Why Try Harder - The Greatest Hits Fatboy Slim Dance 14780691.57
4482 Hard (Explicit Version) Rihanna Soul/R&B/Funk 14780691.57
5394 Wanted Dead Or Alive Bon Jovi Rock 270717.8094
4176 The Climb Various Artists Pop 198067.0679

And sorted by R0 (value used in the publication):

count title artist genre r R0
5273 You’re Never Over Eminem Rap/Hip Hop 2.21E+36 8.35E+38
5039 What Is Love Take That Pop 2.33E+20 5.48E+22
19694 Haven’t Met You Yet Michael Bubl?? Pop 4.21E+18 1.73E+20
5971 Said It All Take That Pop 277136258 80645628008
4252 Trouble Various Artists Pop 7.476801722 2962207752
3944 Reminder (Explicit Album Version) JAY Z Rap/Hip Hop 7.476801722 2962207752
5356 The Captain Biffy Clyro Rock 21745636.89 2849629571
4176 The Climb Various Artists Pop 198067.0679 3203744.967
8677 Silent Night Various Artists Pop 9982.208024 547980.1013
4997 You Take That Pop 1433.370165 335430.6191

And there is a limit to their spread out side of their fans as the ones with the top ten with the most downloads are:

count title artist genre r R0
27801 Bad Romance Lady Gaga Pop 0.433768749 13.0808621
26769 Poker Face Lady Gaga Pop 0.45008816 20.67689975
24270 Paparazzi Lady Gaga Pop 0.966205265 31.07322029
23636 You’ve Got The Love Florence + The Machine Indie/Alternative 0.132649425 1.025016292
23146 Only Girl (In The World) Rihanna Soul/R&B/Funk 0.396186691 14.96233854
23138 Fight For This Love Cheryl Pop 9.677456403 595.2017327
21945 Boom Boom Pow The Black Eyed Peas Pop 0.163211527 24.191656
19694 Haven’t Met You Yet Michael Bublé Pop 4.21E+18 1.73E+20
19516 Just The Way You Are Bruno Mars Soul/R&B/Funk 0.297827675 5.638762101
18517 Just Dance Lady Gaga Pop 0.035876789 1.000680982

Most of which have pretty low r values but still infected the whole population over time. The main exception in the top 10 is Michael Bublé and Haven’t Met You Yet. Ironically a song I have never heard before today. It’s kind of feel good innocuous forgettable pop that will never make the Rolling Stones top 500 songs. Obviously hooked the masses though.

I haven’t tried to reproduce the analysis or go into more depth as far as interpretation goes so there are probably more stories there.

Edited to include R0 values as per publication. Numbers don’t match and I haven’t yet found the exact document that includes the numbers they report.

5 Likes

… this sounds like a job for a Simon and Garfunkel tribute band. Shame I don’t know of any. :rofl:

2 Likes

Weird - but I will take it. I like that album, but so much of their catalog is amazing.

Here is a very rare single that I only first learned about a few years ago. I don’t know why they never re-released it on some obscure single collection. It is just ok, from their earlier rave days,but I like it.

Nice, I didn’t know about their first release and it seems to have been a collectors record at the time too. Thanks for the heads up.

1 Like

Yeah ditto. I’d love to get a copy. Or a CD re-release.

It’s called editing. Sometimes happens post-publication.
:man_shrugging:
:wink:

It is sold as MP3 or FLAC on Bleep.com, but I doubt it will get another re-pressing. I think they did a limited vinyl re-release in the UK in 1997, when all the old Prodigy singles were re-released and they broke some record for having the most singles in the top 75 at once.

Physical copies are very expensive, like in the hundreds of dollars range.

It looks like a lot of the best selling downloads on Bleep right now are by Richard H. Kirk :cry:

3 Likes

1 Like

It’s been called viral marketing for about 20 years for a reason, this is not new news!

2 Likes

Discogs has the full release history. I had left the UK a couple of months before it was released and was discovering German Techno. I used to try and catch up a bit when I returned for holidays but missed a lot. There was a bit of a breabeat scene in German but not much. I do remember the release of Charly and Everybody in the place. Out of Space was big on German MTV too.
I had not heard that Richard H. Kirk had passed, sad news indeed.

That number seems, um, a little on the high side.

Not quite a S&G tribute band, but Scarborough Fair in the Rammstein style works well.

3 Likes

Yeah, I can not recreate table 1 from the publication which is a bar chart of mean R0 for each genre. Only the Bollywood result matches and that’s because it is one song. . . Combine that with the misaligned genre in the top1000GBsongs.tsv file (possibly due to sorting on some columns but not all).

Something is wrong with the data as shared which is a shame as it is quite a high profile and fun paper. I don’t know if I could run the calculations again with the data given and I haven’t looked at the R source code or the other scripts they used to see how easy that might be. I can only think they redid some calculations and did not update the data. I might send them an email pointing that out.

I would suggest another possible reason. Electronica (very infectious) and Dance (not infectious) were mostly the same in Europe in the early to mid 90s, but Electronica ended up being what music journalists liked and Dance was the music that they didn’t.

The infectious stuff is what was heavily promoted, you knew what it was called so you could easily find it. Some of the non-infectious stuff are those songs that you heard in a club at the weekend, you thought was amazing, but because it was played by a DJ and was a white label that meant nobody knew what it was. I am still finding out the names of music I liked in the 90s.

3 Likes

This topic was automatically closed after 5 days. New replies are no longer allowed.