Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Fast DataFrames for Ruby

github.com

Post reply on HN
Print view View original

11–20 of 28 posts

Re: Fast DataFrames for Ruby

#11

Post by goatlover »

claudiug wrote: post #10
oofbey wrote: post #3

I'm trying to imagine why somebody would start a data analytics project in ruby.

I'm trying to imagine why somebody would start writing so hateful comment on hn. how empty the soul of that human can be?

There is a school of thought that you should use the right tool for the job, and some languages are better tools for certain tasks. Python already has all the libraries, and Julia has built-in language support. But nobody is stopping you from using Ruby or JS or PHP for whatever.

Re: Fast DataFrames for Ruby

#16

Post by ipsum2 »

claudiug wrote: post #10
oofbey wrote: post #3

I'm trying to imagine why somebody would start a data analytics project in ruby.

I'm trying to imagine why somebody would start writing so hateful comment on hn. how empty the soul of that human can be?

I didn't read it as hateful, but as legitimate curiosity. Ruby is at a disadvantage because the analytics ecosystem is primarily Python based.

Re: Fast DataFrames for Ruby

#17

Post by swalsh »

oofbey wrote: post #3

I'm trying to imagine why somebody would start a data analytics project in ruby.

Honestly I used Ruby about 6 years ago, but have been a python guy ever since. That said I believe Ruby's main advantage is its metaprogramming capabilities. You can build powerful DSLs in Ruby pretty quickly. Adding analytics to that could be useful in certain cases.

I'd like to see spark bindings first though before I would seriously consider it.

Re: Fast DataFrames for Ruby

#18

Post by pelasaco »

I think its really interesting such gems offering a ruby layer on top of rust libs. One issue that I have with that is - and maybe it's my ignorance - but is that necessary to bundle the original lib as https://github.com/ankane/polars-ruby/tree/master/ext/polars ? I can imagine that makes easier to avoid breakage, or with C extensions, because you dont have some dependency manager around, but couldn't we sort it with Cargo? For instance, having the version locked and it could download (and cache) the dependency when necessary?

Re: Fast DataFrames for Ruby

#19

Post by petepete »

ssaunier_ wrote: post #12

ankane’s gems are truly something.

Blazer is a particular favourite of mine. A former colleague taught herself SQL from zero knowledge by looking at and piecing together bits of other reports, experimenting with familiar (interesting) data and going on to build dashboards her team loved.

If we'd have just given the team a locked down Data Studio/PowerBI report none of that would have happened. Encouraging people to peek under the hood can be a huge benefit.

Re: Fast DataFrames for Ruby

#20

Post by vidarh »

swalsh wrote: post #17
oofbey wrote: post #3

I'm trying to imagine why somebody would start a data analytics project in ruby.

Honestly I used Ruby about 6 years ago, but have been a python guy ever since. That said I believe Ruby's main advantage is its metaprogramming capabilities. You can build powerful DSLs in Ruby pretty quickly. Adding analytics to that could be useful in certain cases. I'd like to see spark bindings first though before I would seriously consider it.

http://ondra-m.github.io/ruby-spark/

https://github.com/ondra-m/ruby-spark

Hasn't been updated in a long time - no idea if that's because it's complete enough or if it's been abandoned, as I don't use Spark so I haven't tested this gem myself.

Post reply on HN