R v. Python

I'm a stats/computer science student looking at a career in ML or """""""""DATA SCIENCE""""""""""""".
Which would be more worth my while to master?

both are good languages. Honestly it depends on what you're after. Much of the statistics side of the biological sciences are dominated by R. It has a very robust statistical package and easy to use user interface (if you download Rstudios i believe). That being said, R is a very high level language. As such, you lose a lot of the flexibility and computational prowess lower level languages often have.

Python on the other hand has some really great libraries like numpy, scipy, and sympy, among many more. Python is a lower level language than R and so you may often times find yourself doing many more tedious actions. Granted, this comes with greater flexibility.

Python is said to be an excellent 'glue' language. It is capable of bringing together many different functions written in other languages and complete and overarching task well. It is not uncommon for Biological modellers to use fotran or C for highly computationally taxing operations and embed it into the python language. This tactic combines the efficiency of low level computational power with high level easy of writing.

So what should you invest your time into? It doesn't really matter. Learning a stats package in general is very helpful. R may take less time. Or just learn Mathematica.

Fuck off, one more fucker. It is already hard to get a job because of people like you.

Don't know about R but most people I know use python

Fuck python

Python is pretty much the standard in data science nowadays, I believe. It has an abundance of data science libraries, and it's incredibly easy to learn.

RStudio is shit. Use RGui.

Just git gud.

Python isn't really low level, I like your glue analogy better where python can be used for various purposes and has a wide range of libraries that can make learning it valuable for many different applications, R on the other hand is good solely for data science and stats.

I like the *idea* of R, but Python seems much morr comfortable to me for some reason--despite professors and textbooks pushing R so hard.

Thx

Piggybacking off this thread; all data sci books use R. Can you guys recommend any that use python (more comfortable using python)

There both complete shit, but python is by far the least shit of the 2.

python _is_ shit. Its a better bash script, treat it accordingly,

python:
>can do everything
R:
>can only do math

hmmm I wonder which one is better to master...............

fuck that CS shit go full stats my guy, ML is a meme

Scientists generally don't need all the extra features Python offers and R excels over Python for the things it's more narrowly focused on.

Also the real answer is to learn both, they're both piss easy languages.

use julia, brainlet

python is just glue.
there's nothing really profound to get at within python. learning python is like practice for programming in general. there are lots of tools that you can interface via python, but you'd be working with the tool, not some built-in greatness of python. So you could say, let's use NumPy for statistics, but you'll soon find out you want to combine it with other tools out there and things just get more obscure. R is going to be the better all-around statistical tool because it is designed to be that way. Considering that you're already going to be doing programming in other languages, R also gets you broader experience.

I do research in computational chemistry, and I use python basically as you described. My heavy chemistry packages are almost entirely in C++, but I use python as a top-layer shell to operate them, parse and organize data, etc.

For my purpose, I doubt I will need any languages other than C++, Python, and CUDA if that counts.

why

>be scientist
>need to do some stuff quick
>do a for loop
>takes 100 years in R
Python is just way better if you need to do any scripting at all, which you probably will if e.g. you need to change all your data to a different format by hand. Have fun doing that in R.

data science is the biggest meme I ever fell for since "go to college lmao"
I'm on the brink of suicide.

learn both, they are piss easy. Literally meme tier languages. It's not like you will have to commit half your lifespan and your first born's eyes to master like C++

>do a loop
>in R
Well there's your problem

Also what the fuck are you talking about. Wrangling data is one of the things R is actually good at.

Same here, kiddo. Fucking hypers. I will probably kill myself this summer.
> LMAO data science is in demand dude

I'm just dropping out of stats next semester. Data"science" is a meme and a boring one.

Are you going neet?

>R v. Python
Julia >>> R & Python

Damn. I was entertaining the possibility of getting into data science. Are you guys in the US? Why is it such a bad choice? Is it only the lack of jobs? Are there really no future prospects to it?

HAHAHAHAHHAHAHA fuck dude I almost poppe a vein there

I am from the UK. It is actually quite interesting, but it looks like you either need a PhD or to be a black trans gay to find a job.

how to put it... you and 100k more like you are going to apply for the same job.
That's how memes work, by the time the news get to you it's already late to make money in that field. Ever heard "buy the rumor sell the news"? Rumor was years ago, nowadays these fuckers are selling the news, to you.

Julia is /ourfuture/

as fast as python to prototype
but close to gcc to run.

t. perfection.

MATLAB unironically

depends on where you look. if you're a trained bench scientist and also know how to do your own data analysis, you're gonna get a job really easily. if all you know how to do is program and have no other skills, you're shit out of luck

Viral, go take a shit in the street

> if you're a trained bench scientist
ie has a PhD.
> and also know how to do your own data analysis
Doesn't worth shit without PhD.

No idea about R but I really like python,many cool features for analyzing data.

It's a pretty big problem if you're a scientist and you just need to write a quick script.
Maybe if you can use packages with inbuilt convert functions but if you need to convert a file to some stupid relatively obscure format to use it as input to some program, you choose python. It's just more flexible.

>It's a pretty big problem if you're a scientist and you just need to write a quick script.
Almost every native function is built to accept vectors as input. Learning how to write efficient R code is literally as easy as putting your input into a vector.

It's not a problem at all. It takes five minutes to learn how to fix.

And no, you're just wrong. Unless you're working with some shitty proprietary data format, a huge amount of scientific programs out there (especially the ones you typically turn to R to analyze) rely on flat text files for their data input/output. R reads/writes delimited text files like nothing else.

>Almost every native function is built to accept vectors as input.
Doesn't help if I just want to do some quick if else in a for loop. Why should I go through some weird unintuitive sapply trickery when it's just much easier in Python?
>R reads/writes delimited text files like nothing else.
What if you need to combine information from several files to make a file in a specific format? Python is just easier to use in most cases.

>Why should I go through some weird unintuitive sapply trickery when it's just much easier in Python?
Most of the time it's not sapply trickery at all. You just pass your function a vector.

>What if you need to combine information from several files to make a file in a specific format?
You read in multiple files, you build a data frame, you write a data file. It takes a handful of lines.

Get a masters in a STEM field and be able to actually TALK those who are in charge of hiring. You're dealing with data which has been collected on other humans, how can you possibly be good at your job if you can't even hold a conversation with a Chad from management or a Stacy from HR. If you're pic related, become the exact opposite.

Python.

Versatile as fuck, clean syntax, rich libraries, has applications in almost everything

R is completely useless for anything other than pure statistical modeling and it's shit at that

*to those