/sci/ - Science & Math

>>10632645
first learn python u nig

>>10632645
>why not learn R?
It doesn't scale past like 1 million data points

>>10632645
>Flexibile `n' Fun
delet this you fucking jigaboo

>>10632645
>data science
Fuck off, we're full.

>>10632858
where to i start to learn python when i dont have a project to necessarily learn it for

im a recent engineer grad, learned some c++ and matlab

>>10632645
Learn J

>>10632645

its a meme lol its a huge fucking meme filled with females that can't into anything else

>>10632645
data “science”

>>10633195
> As if anybody who data mines for a living doesn't want to kill themselves over their meaningless existence

>>10632645
Linear algebra and derivatives are the basis of neural networks. You don't need to know much more than dot products though, and Tensorflow handles backprop (derivatives) for you. I'd say stats is more important.

Don't learn R, learn Python and Python's DS libraries. Python can do everything R can do and is extrapolable to other fields.

anyone using Jupyter? is it actually worth the hype?

R is a horrible choice. Learn Python.

>>10634580
>>10634460
t. compsci brainlets

actual statistician here, R is way better for knowing what you are actually analyzing.

>>10634583
Care to elaborate? Why? How?

>>10634586
Not him but R was specifically made for shit like statistics and has a lot of packages out of the box for it

>>10634594
>>10634586
Yep, and you need to understand statistics in order to use R.

>>10633229
Learn python the hard way

>>10632645
this >>10632858 then learn SQL you nig

Word of advice, don't.
Waste of time. Something to be left for the H1B drones.
Study medicine, take the MCAT, become a racist and homophobic physician. Or a priest. That is the final redpill.

>>10632645
i hate R so much, the only reason its good is the amount of packages

>>10632645

Learn to be really good and efficient in Excel.

Whatever fancy programming, database or CRM you use, you will eventually need to export all that shit in Excel.

80% of your workload will be messing around in Excel. The people viewing your data will be normies who will insist on getting it as a spreadsheet. And most of your work will be in meetings with them going back and forth over an excel spreadsheet.

I have worked in this profession for years and participated in projects to launch the modern looking online reports and visualisation tools. All the higher ups I work with ignore them and ask for the spreadsheet.

>>10634976
Whats there to hate?

t. Rpro

>>10634995
tfw I was literally told to drop excel because sql is a much better thing to have

>>10634886
/thread.
fuck bitchboy jobs that command zero respect and you can be fired in an instant. muh learn SQL. muh Learn R. muh Learn programming. meanwhile, a billion Indians are at the gates. meanwhile you need to fill out diversity forms. friends a doctor, he applies a place, and he gets a call back within 24 hours. no diversity forms. doesnt need prove himself with 5 interviews, a panel of tests and all this bullshit. no non-competes. even a pharmacist is higher up the rung than data 'scientists'.

>>10635026
>>10634886
>/thread.
>fuck bitchboy jobs that command zero respect and you can be fired in an instant. muh learn SQL. muh Learn R. muh Learn programming. meanwhile, a billion Indians are at the gates. meanwhile you need to fill out diversity forms. friends a doctor, he applies a place, and he gets a call back within 24 hours. no diversity forms. doesnt need prove himself with 5 interviews, a panel of tests and all this bullshit. no non-competes. even a pharmacist is higher up the rung than data 'scientists'.
You think pajeets are intelligent enough to actually do statistics, not some python shit kaggle tiet graphs, sissy boy?

>>10635024

In your real job, your manager or whoever you are reporting to, there is a 90% chance they won’t even know what SQL is.

They will ask you for a report. And you will spend 80% of your time exporting your data into Excel, making charts for them, going back and forth via email and attending meetings.

It is essential that you know your way around whatever system the company uses to store the data. But excel will still be your bread and butter.

>>10635039
Enough to fool the low IQ people hiring them.

>>10635024
You got memed son. Big time.

>>10632645
As a guy with a major in data science. I'll say this much. Most computer science and "'""'data science""""" theory is easy enough to work through if you have a strong math background and a degree in a qualitative field (a real qualitative field like math, engineering, physics not economics). You might want to minor in CS if it isn't more than an extra year of study or alternatively minor in statistics or double major in that (which is frankly better). This will give you a solid foundation for interpreting data. The actual tools (like machine learning and neural networks) aren't too hard to understand. Again, very easy for someone like you. But absolutely do linear algebra if you can. If there is like a 'Linear Algebra for Computer Science' or equivalent and a regular 'Linear Algebra'. Choose the latter as the former has been dumbed down for CS students.

Don't start with R. All the libraries using in data science like matplotlib, Tensorflow, etc are in Python. Learn Python, and for anything high performance learn C++. Assuming you're doing computer engineering you've probably already done C or C++ (or both).

Discrete math is also very cool. Again avoid the CS version and go for the raw version. This is much harder, and if you haven't done many proofs you will definitely need to brush on your knowledge of proofs. There are many.

>>10635024
I need to write plenty of queries on databases and then I do the finishing calculations in excel. Excel has power query which is really nice after your initial SQL to get exactly what you want

>>10634995
>>10635024
Friendly reminder most of the global financial system runs on Excel. Let's that sink in.

>>10634583
I'm the second guy and I'm an EE. Python is a versatile language that you can use in any field. If OP ever gets bored of doing data science and wants to switch to, say, computer vision, knowing R will be almost worthless, whereas knowing Python will be useful. I would never recommend a student to learn a domain-specific language before a general-purpose language.

And again, there's nothing R can do that Python can't if you know how to do it.

>>10634460
>Python can do everything R can do and is extrapolable to other fields.
python is kindof shitty for statistics if you wanna go beyond running a premade algo on a problem

>>10635520
And R is kinda shitty for general programming tasks if you wanna go beyond just running statistics packages on a data set

>>10635556
Quite simple, Python for production and R for statistical analysis.

>>10635381
>Let's that sink in.

>>10635173
> a degree in a qualitative field (a real qualitative field like math, engineering, physics not economics)
> qualitative
I think that you meant “quantitative”, but feel free to correct me if I'm wrong.

>>10635587
Wait, what the fuck happened with my pic of Inigo Montoya? Does anybody know why the thumbnail is different?

>>10635586
Kill yourself you waste of space.

>>10635590
You uploaded it in the same microsecond some other guy uploaded those titty laidies. He must be asking why is Inigo Montoya in his thumbnail.

>>10635600
Thank you for your explanation. You deserved those dubs.

>>10635600
Moots ghost strikes again

>>10635587
I did. Thanks for the correction.