Short answer to the original question (already answered many times) is different music, different muscians, different processes, different eras in musical history, different budgets, different engineers, etc. etc. etc.
For me, I do my best to use the least amount of tracks that the music I’m making requires. However, I have no qualms about adding more tracks if something I want to do would make sense, or sound better on its own track.
If say, I’m mostly using an 808, I’ll cram the kick, snare, and maybe a few other sounds into one channel (two if it’s not really an 808 
) and then maybe I want to add some space to the sound. I’ll toss the hats, shaker, clave, and other higher frequency sounds on their own tracks with their own effects, and maybe widen the field a bit with some stereo effects. Bass sounds always go on their own track, so I can EQ them against the lower frequency percussion. Sometime I do the same with the kick, but not always, as it doesn’t always matter. Maybe some modular sound effects, or something semi-melodic. Each instrument goes on either a mono or stereo track. So that’s probably pretty close to 8 tracks or so.
On the other hand, sometimes I want to make a stereo track with just my modular, using serial effects. So, two tracks.
If I’m writing music with software, I usually end up with 12-24 tracks or so. The reason being, that I have little hacks I do for certain things. Like having four 303 tracks for a single bassline. The reasoning here is that I don’t have to program pattern changes. I can just swap between tracks to change patterns. Then I can also tweak them differently with automation too, so it’s just easier for me to bounce back and forth that way, or compare two parts, instead of paging back and forth, I can see both. I tend to program more intricately in software too, so I like lots of little sounds and tiny effects happening, where more tracks is very helpful. I don’t do multitimbral arrangements on one synth (in software at least, with Elektrons I obviously do) so having a few copies of the same synth with different sounds is vastly easier than doing it all within the synth.
Anyway, just a few reasons that I personally have for various track counts.
I write music much differently depending on which subset of gear, or hardware vs software, etc. I’m using. With software, I program every single note, automate ever single change, and create something very structured. If I’m using hardware, I generally do this more for fun, or for live playing/recording, so the music is less intricate in the way the software stuff is, but it’s more organic and human sounding, because I’m doing everything on the fly. Other times, like the modular example, I’m just plain experimenting or designing sounds, and sometimes hit record.