Spotify's Release Radar playlist is an excellent weekly reminder to do your data analysis using the [unique item ID] field, not the [item name] field.

(restarting thread because it kept tagging accounts and I can't untag them)
This week's example: this is not the Brand New that I used to love but find hard to listen to now after hearing about the singer's behaviour. This is Brazilian (I think) hip hop. It's very different.
This week's example: this *still* isn't Basement.
This week's 2nd example: I think what's happened is that I've been on a 90s emo/indie/early post rock kick lately (The Appleseed Cast, envy, etc.), and they're recommending me Tortoise, who I don't know but seem to be that kind of band. But, this is definitely the wrong Tortoise.
This week's example: this is, again, not the abuser-led goth/punk/emo band Aiden that I'm embarrassed to say I loved as a teenager. This is a rapper with the same name. They're not the same thing.
This week's second example: this is not UK pop-punk kids Martha. This is some simple keyboard lines that sound a bit like The Sims. It's quite different.
This week's example: yet again, not the abuser-led goth/punk/emo Aiden I was embarrassingly into as a teenager. This time, it's not a rapper, it's some generic calm piano music. Either way, it's wrong.
This week's second example: this Microwave is not the shouty indie alt rock Microwave I listen to. It's, I dunno, house ish? I've got a shockingly poor knowledge of genres without guitars. Point is, it's the wrong artist again.
This week's example: is this +44 the one that's 2/3 of blink-182? In their own words: No, It Isn't.
This week's example: this is not metal band Baroness. This is ... electro-ish? house-ish? I dunno, it sounds like clubs I feel out of place at.
(just realised I should have put a link to the old thread right at the start. it's here:) https://twitter.com/GwilymLockwood/status/1101562284139266049
This week's example: it's a reappearance of Martha and Freaky DJs. This Martha still isn't the UK DIY pop-punkers I sometimes listen to.
This week's example: it's Martha and Freaky DJs again! Even though I keep saying I don't like Freaky DJs!

Algorithm: hey you listened to a teenage nostalgia song once, here, have all the The Used forever

Also algorithm: you sure you don't like this? Nah, you definitely do
(similarly, I tried out a colleague's trance work playlist for about half an hour once. It's nice to try new things, but it's not for me. I've had this in my release radar every single week for months. I say "I don't like Above & Beyond" every single week. It does nothing.)
(seriously, just do like:

AllTheData[AllTheData['NoThanksFlag']==0]

before doing the algorithms, that's it)
This week's second example: this isn't Belgian hazy bois Newmoon. This is Iranian (?) rap.
tfw an algorithm looking for spotify content misinterprets your thread moaning about spotify algorithms
This week's example, pt. 3: this is not Cali post hardcore band Movements. This is a rap artist with the same name. And look, even Movements know about it!
This week's example: unless this is a confusing comeback from former Cambridge alt rockers Mallory Knox, we've got another [Artist Name] / [Artist ID] mix up here.
This week's second example: it's the rapper Citizen again, not the grunge-flavoured post-hardcore band.
This week's third example: this isn't VLMV. This is by somebody called El Ultimo Almanauta, because I found it on youtube. VLMV changed their name from Alma to avoid this kind of thing. I don't understand how this is still happening.
This week's example: another week, another Freaky DJs ft. Martha song that isn't Martha the UK pop-punk band, another "I don't like Freaky DJs" action that hasn't done anything.
This week's second example: this isn't Florida emo boys Capstan. This is Italian rap. It's not bad, but it's not right either.
This week's example: no idea what this mm-tss mm-tss mm-tss noise is, but it's definitely not the metalcore Architects I listen to a lot.
This week's second example: again, this isn't pop-punk Martha, but it's not the Freaky DJs and Martha combo either this time, which is progress, I guess? Still wrong though.
This week's example: it's been a couple of weeks without a Martha that isn't the pop-punk Martha, but we're back with a bang today. More things on Spotify should have clip art in the image.
This week's second example: that's not that Brand New. Also, oof.
This week's third example: that's not that Aiden. Also, OOOFFFFF.
This week's example: death, taxes, and Freaky DJs feat. Martha in my recommendations.
This week's example: this does not feat. slow sad post rock band VLMV, which is pronounced "alma" but spelled VLMV. This is a different string which is pronounced the same. That lookup/join shouldn't be happening (and shouldn't happen anyway, because you should use unique IDs).
In an entertaining turn of events, this has now happened to my own band.
This week's example: it's in my release radar even though I've taken steps to get it reassigned! This The Pangolins is not *my own band* The Pangolins. This The Pangolins is a totally different The Pangolins.
This week's example: this is not the Aiden from my teens I'd rather forget. It's something totally different. Bonus points for getting the wrong artist and sticking them in the playlist cover image.
This week's second example: this track doesn't even have Owen listed as an artist, it's just on an EP that features Owen. Either way, pretty sure it's not Mike Kinsella's Owen.
This week's third example: this Loathe is not the Liverpudlian metalcore-ish post-ish band.
This week's example: it's the wrong Aiden again.
This week's example: another rap Owen that isn't Mike Kinsella.
This week's example: this isn't Japanese screamo/post-hardcore stalwarts envy. This is ... just terrible.
This week's example: is this new music from Finnish love metallers HIM? No, of course it's not.
This week's example: this is not defunct New Hampshire emo boys Old Gray. This is something in Spanish, spelled Old Grey but miscategorised as Old Gray, with a logo from a barber's.
This week's second example: this is another Owen that isn't Owen. This is a Russian (I think?) rap about TikTok. It's very different from the sad waltzes of Pietro Crespi.
This week's third example: it's been almost three months since the last time the wrong Martha turned up in Release Radar, but here we are again. This is Italian rap/hip hop (I dunno, I'm not good at genres), not North East English pop punk.
This week's example: again, not HIM.
This week's example: this isn't Philly twinkly emo kids Marietta. This is some Brazilian soft-rock / pop / rnb-ish thing.
This week's second example: another envy that isn't Japanese post-hardcore envy.
This week's example: again, this is not that envy.
This week's second example: I think it's tried to suggest PEARS to me, and ended up with this? It's almost definitely not what it's meant to be.
This week's example: this is not Austro-Brit crusty emos Morrow. This is rap.
This week's example: this is not post rock / emo Gates.
This week's second example: it's been a while since I've seen this one, but this is not rock / pop-punk Transit.
This week's example: this is not Aussie indie band Ceres. This is rap in Spanish.
This week's second example: this is, again, not Transit. But it's not rap this time, it's some kind of funk / pop. Still not Transit though.
This week's third example: I don't know how this got in there. My theory is that Spotify has mistakenly given me obscure rap / hip-hop artists with names of bands I like so often that it now thinks I like it.
I'm looking forward to seeing how this works when Spotify's recommendation algorithm can't tell the difference between artists with the same name. https://twitter.com/thefader/status/1323340529883643908
This week's example: this chill out instrumentals Cursive is not the famously not-very-chill Cursive I normally listen to.
This week's example: this autotuned rap Basement is not Suffolk sad boys Basement.
This week's example: again, wrong Basement.
This week's example: not Japanese, not screamo/post-hardcore, not envy.
You can follow @GwilymLockwood.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled:

By continuing to use the site, you are consenting to the use of cookies as explained in our Cookie Policy to improve your experience.