Python began to pick up momentum around 2000. It was able to break out of the "scripting" niche because it had good OO support and encouraged a consistent style that made it easy for several developers to work on a large project.
Rails probably marked the tipping point, but it wasn't just the framework. A lot of Perl programmers found out they liked Ruby. You're right that Perl 6 was part of the problem; it might be where Ruby is today if it had come out in 2003.
On the other hand, Ruby probably wouldn't exist if Perl 5 had a better object system.[1] It's "Ruby on Rails" because DHH fell in love with Ruby.[2] The creators of Django learned Perl before falling in love with Python.[3] So that has something to do with the languages too.
[1] http://www.linuxdevcenter.com/pub/a/linux/2001/11/29/ruby.ht...
[2] https://www.quora.com/Why-did-David-Heinemeier-Hansson-choos...
[3] http://www.akitaonrails.com/2008/01/01/chatting-with-adrian-...
Both are false. Data science libraries are quite new development that happened because Python got popular, not the other way around. Backing by Google (whatever you mean) was irrelevant, as the main selling points for Python dozen years ago was simple and readable syntax and nice web toolkits (a movement that followed Ruby with its RoR).
> If Perl6 hadn't taken 15 years to complete Perl's popularity might be different.
Perl6 didn't compete for popularity with Perl, because it was essentially non-existent back then. Perl popularity decline cannot be attributed to Perl6.
The thing is, Perl is a very complex language, but has competiton that is simpler grammar-wise and similarly powerful and expressive: Python and Ruby. It's easier for newcommers to learn Python or Ruby than Perl, so they rarely dig into Perl. This means that the number of people leaving the language (a normal thing for any language) is not matched by the number of young programmers, which leads straight to decline.