You have just saved a considerable amount of time and nerves. Not only will they fundamentally change it in version 2.0, they are going to use a different language - TypeScript flavour.
Do you have any domain knowledge outside of programming? A data scientist who understands the domain they're dealing with is a rarity and at least in my opinion worth more than someone who knows all the math, but has no grounding in what he's actually trying to analyse.
ECMA-262: around 250 pages, most of which is just water. One day reading. No new concepts are introduced over anything that existed 30 years ago, so it does not require any learning, just knowing where to look for a reference.
> client side frameworks
Who need to learn the libraries? They must have some documentation, I suppose. And again, I really doubt any of those managed to introduce a fundamentally new concept which was not known 30 years ago.
> but haven't monitored the developments of the language and best practices for the last 3 years
The safest assumption is that nothing really changed.
> I like data science
I.e., statistics, a > 100 years old science.
I am not saying it would be difficult. But I lack the energy. I have read a lot of documentation in my 30 years of programming.
I used to be a Java developer, but haven't monitored the developments of the language and best practices for the last 3 years. Also I don't really like Java anymore.
I ponder getting into Android, but it's Java again, and then I find I have to learn a whole new style language that is almost but not quite like CSS, but seemingly less powerful.
I don't really care about agile development or scrum, another deal breaker for many jobs.
I like data science, but it seems mostly PhDs fresh from university or other young people are being hired for that. I actually enjoy doing Coursera or Udacity courses in data science, but it's not enough.