does statistical programming belong on /g/ or /sci/?
/g/
we need more competent people in /dpt/, there's too many fizzbuzz babbies who do nothing but argue about syntax
>>51836555
cool cool. I'll freely admit to not being a real programmer, just a biologist who got more into script-based analyses than my peers and realized that knowing how to write a for-loop would help me more getting into grad school than having spent four months on a deserted island counting penguin shits.
If you use R, sci
If you develops functions for R, g
>>51836588
>not using apply
>>51836625
loops are more universal, which is why I chose to use that as an analogue for 'having some level of basic programming knowledge'
>>51836539
>does statistical programming belong on /g/ or /sci/?
I don't think it matters.
Chances are that you won't get quality responses to the topics you are interested in on either board.
>>51836539
>using a sub-Python-tier language
faggot
>>51836675
fair enough
>>51836691
could be worse. could be matlab.
>>51836539
lurk both post in threads that scrape what you're into
>>51836654
Yes but they are horrible implemented in R.
It is a great tool for statistics and you can make anything you want with the language, it will just run very slow.
I also appreciate the syntax and that is is not that different from c++.
And documentation is really good.
>>51836733
all true
I, unfortunately, learned loops early on and am working on breaking that habit.
but yeah, documentation is really good (TY Based Hadley)
>>51836713
could be worse. could be mathematica
>>51836819
Mathematica is great if you only deal with symbolic math, but as a programming language, it is terrible.
>>51836539
I am not going to bash R OP but do consider learning python. Being able to code in python is very important for what sounds like an ecology-esque biology major (based on counting penguin shits). You can model populations and shit.
God I hate R.
But, if you really must use it, read "The Art of R Programming". It provides a fairly good foundation in that crappy language.
>>51836897
that's my planned next step, in a couple months when work pace dies down, gonna start by trying to convert a bunch of stuff I've done in R to Python and go from there.
>>51836897
Relevant: http://austringer.net/wp/index.php/2012/10/08/population-modeling-in-python/
>>51836913
i have heard good things about the biopython library
>>51836932
did Leslie matrices in R a year or so ago
>>51836950
I do ecological stuff, so population models, that kind of thing. biopython is for molecular stuff. I'd mainly be working with numpy/scipy/pandas.
>>51836539
>>51836588
Biologist? You should look into genetic programming and symbolic regression on R. I am a civil engineer but found great use for it. It basically automatically finds a mathematical formula that describes your data. Say you have a 'y' value and several 'x' values. Then you can automatically find y = f(x1,x2, x3....xn). It makes data analysis simpler when there are tons of parameters involved like it usually is in biology. I don't use the R packages for it, I use Eureqa,. But I know there are R packages
>>51836979
I'm trying to learn to work in the Bayesian framework, but I don't have the bandwidth to learn the necessary math while in end-of-year crunch.
Undergrad EE with lots of freetime. I know how to program, we are forced to learn MATLAB, would R be worth learning?
>>51836588
>using loops in R
Top kek, gtfo
>>51837946
If places you're likely to work won't have licenses for matlab, then maybe.
>>51836539
nothing intellectual belongs on this board
I'm learning R currently as well though for muh meem major
>>51837946
I don't see the parallels between EE and stats... I don't know too much about the range of things that could be represented in R. I thought it was a language that was strictly focused on data manipulation and statistics
>>51843476
meet your new best friend
http://adv-r.had.co.nz/
>>51844276
Thank you. I wasn't aware of this site mostly because I've been using different resources and sticking to them. Will see what knowledge I can reinforce with it
>>51836897
Python is good for some things, like machine learning and web scrapping.
For things that require rigorous or tradicional statistical models, like inference, R is probably a better choice.
>>51844353
Hadley has written more packages that I use regularly than basically anyone else.
Read his paper on tidy data, it'll change your life.
>>51844514
Python has matplotlib which make prettier graphics than R does by default.
>>51844626
Interesting. This is a topic I've been meaning to dive into. Thanks for the resources. I'm much less informed about 'data science' than I am about traditional CS concepts
R is great.
I'm working with traffic GPS data right now that was web scrapped, works like a charm
>>51844951
Ever heard of ggplot2?
Is literally the best tool for graphics right now, the python community even tried to replicate the package. It's still in testing phase though, but you can get that in http://ggplot.yhathq.com/ if you're interested
>>51836979
>you can automatically find y = f(x1,x2, x3....xn). It makes data analysis simpler when there are tons of parameters involved like it usually is in biology
lots of languages give you a polyfit function bro
>>51836819
there isn't a problem with mathematica if you're working symbolically
>>51837946
Either use an open source alternative to matlab (eg. octave) or be sure that you'll need stats on your works, that is, if you're into Communications, fading models and whatnot. Otherwise another language might be far more effective
>>51845165
that's a pretty desktop
>>51836979
Genetic programming has 0 things to do with biology and has also not been used since the 40's because it's absolute fucking trash.
>>>/reddit/ with your fucking memeshit.
>>51837946
python
>>51845949
>Genetic programming has 0 things to do with biology and has also not been used since the 40's because it's absolute fucking trash.
It's used in ALL kinds of fields that involves data. It is great for chemistry to find cool correlations. I recently used it to find correlations of Phosphorous in wastewater and hope to learn cool stuff about it. POSSIBLY even a way to calculate or estimate phosphorous without taking samples to the lab.
It might be used in biology to find growth patterns and to calculate the needs to grow certain bacteria. What temperature, feed and what yield you can get in these or these conditions. And it is often used in bio informatics research, but it's kind of specialized. But it is super useful.
You are talking out of your ass
>>51847694
Confirmed for elementary school kid. 4chan is for adults. Leave.
>>51845193
Yes, the difference is that it does not assume any "form" it checks all sorts of connections and might reveal something better than normal linear or non-linear regression.
>>51847722
>I am clinically retarded and there's nothing you can do to stop me!
>>51836979
The new trend now is to use Probabilistic Graph Models for inference, such as Bayesian Networks like pic related for some problems
>>51847713
I use genetic programming in my own research and was just giving someone advice about some software that will make the job easier. Stop being a asshole.
>>51848042
>I use tech nobody's been using for over 60 years because it has been deprecated a billion times over
>therefore it's alive and well!!11
Literally worse than women's degree tards.
>>51848011
Which is why everyone should learn STAN (or venture, or maybe church).
>>51836539
Statistically speaking, it belongs on /sci/
Speaking in terms of programming. It belongs on /g/
R is plenty fast if you're not dumb and you don't do dumb things like use loops instead of functions that use inherently vectorized compiled C libraries under the hood
>ggplot2
>all those packages on CRAN
>indexing second only to MATLAB in ease of use
>functions print their source when you call their name
>fast if you know what you're doing
>all those stats functions
>mfw
>>51848292
bestplot2
>>51848187
That is stupid, that is like saying nobody is using arithmetic anymore. Math is math and it's a great way to get solutions and it is one of the most popular algorithms in research.
On google scholar there are 43 300 papers on "genetic programming" from 2015, but only 17 300 on "Probabilistic Graph Models"
Genetic programming is much more in use. You are just looking up random stuff on wikipedia and pretending to know stuff. It makes you look like an idiot in front of anyone who actually know math.
Probabilistic Graph models don't even give results in a similar form. You should get a science education, kid. Many of us have
>>51848555
It's like saying that nobody uses a horse to go places nowadays because cars completely deprecated them because that's literally fact.
>>>/out/ with your worse-than-women's-studies toiletpaper ``expertise''
>>51848627
Genetic programming is better, which is why it is used more in contemporary scientific research. Great for symbolic regression and easy to use.
You need to learn some math
>>51848671
5/10, I thought you were serious for a while.
