Anyone use R?

[Endoxifen]

---

Members do not see this ad.

Does anyone here use R for statistics? How opaque is it for a pre-med to learn? Thanks!

 

[flapjack3d]

There was a thread about this on Allo a week ago. You could look up what was said there. In a few words, it doesn't really matter what you use, generally speaking. They all have their advantages and disadvantages. Tbh I'm sure excel would suffice for the majority of med student's and premed's projects. I am partial to SAS because it's what is most common in the medical world (you can download University Edition for free!) and SPSS because it's a breeze to learn. If you're thinking of doing engineering or hard sciences research, then go ahead and learn R.

 

[Lucca]

R is not hard. Straightforward. Lookup the package you need to get what you want done and use it. Best for working with multivariate datasets that have already been collected.

To do theoretical work with R you can use Mathematica to produce tables of values produced by some functions, export as csv, and then use R to visualize more powerfully and intuitively than Mathematica's internal visualization software. M has the benefit of being able to produce animations when working with single variables though.

 

[Yuanjing]

Anything you do that is related to statistical models and inferences will make R a handy tool regardless of profession.

That said, I hate R with all my guts.

 

[Endoxifen]

Thanks everyone! @flapjack3d, I'll definitely look up the thread in Allo. Before this I was using Minitab, but I need to make ROC curves which Minitab can do without a plugin, which I can't get because I'm working off a university computer. I also need to use Bayesian likelihood ratios, which I can't do on Minitab easily either (apparently).

 

[honey0102]

Does anyone here use R for statistics? How opaque is it for a pre-med to learn? Thanks!

I used it. There are online videos and such that you can use to teach yourself. See if your university has stats workshops. Mine did and it also had a stats walk-in room where you could get help.

 

[Ehwic]

There was a thread about this on Allo a week ago. You could look up what was said there. In a few words, it doesn't really matter what you use, generally speaking. They all have their advantages and disadvantages. Tbh I'm sure excel would suffice for the majority of med student's and premed's projects. I am partial to SAS because it's what is most common in the medical world (you can download University Edition for free!) and SPSS because it's a breeze to learn. If you're thinking of doing engineering or hard sciences research, then go ahead and learn R.

Is it really true that SAS dominates the medical research world? I always presumed R would be given it's cost and statistical abilities.

 

[LePepe]

Try Swirl for R

 

[libertyyne]

Try Swirl for R

This is excellent advice. It walks you through all the major functions in actual R. There is a JHU course on coursera or ed x about R as well if you are so inclined.

 

[JB50]

I typically use PG13 because my parents set the parental settings as such. Little do they know I put malware on their computer and found out the passcode so I've been staying up past 9:30pm watching R's for years.

 

[Endoxifen]

Is it really true that SAS dominates the medical research world? I always presumed R would be given it's cost and statistical abilities.

My PI (a neuro-oncologist) was the one who suggested that I learn R, not SAS. He said that R is likely to become the de facto medical statistics software in the coming years. He still uses SAS for a lot of things, but he suspects that R will win out in the end. I can't really comment, but he seems to know what he's talking about 🙂.

 

[Medic741]

+1,000 for R --> have heard great things about Swirl as well for learning the software.

Also... if you get into big data analysis R will be pretty much mandatory to know how to use. Eg. Metagenomics/genomics have a lot of R packages that are enormously helpful.

That and time series analysis is really easy to do with R tools.


Very easy to learn with the variety of online tutorials available. My advice... find a few tutorial, and follow them through to the end.

 

[Endoxifen]

+1,000 for R --> have heard great things about Swirl as well for learning the software.

Also... if you get into big data analysis R will be pretty much mandatory to know how to use. Eg. Metagenomics/genomics have a lot of R packages that are enormously helpful.

That and time series analysis is really easy to do with R tools.


Very easy to learn with the variety of online tutorials available. My advice... find a few tutorial, and follow them through to the end.

I've been told that python is a must know for bioinformatics. How different is coding in R and python?

 

[LePepe]

I've been told that python is a must know for bioinformatics. How different is coding in R and python?

Pretty different tbh. R is unlike most other programming languages.

 

[Endoxifen]

Pretty different tbh. R is unlike most other programming languages.

Do you know why you'd use python over R? It seems like R is best for the bulk of analysis, while python may give you novel ways to organizing and differentiate between data sets.

Edit: typos... many typos

 

[LePepe]

Do you know why you'd use python over R? It seems like R is best for the bulk of analysis, while python may give you novel ways to organizing and differentiate between data sets.

Edit: typos... many typos

I'm not a comp sci major or anything, but here's what I found in terms of Python vs R for data analysis:

In Python, sklearn is the “primary” machine learning package, and pandas is the “primary” data analysis package. This makes it easy to know how to accomplish a task, but also means that a lot of specialized techniques aren’t possible.

R, on the other hand, has hundreds of packages and ways to accomplish things. Although there’s generally an accepted way to accomplish things, the lines between base R, packages, and the tidyverse can be fuzzy for inexperienced folks.

source: https://www.quora.com/Which-is-better-for-data-analysis-R-or-Python

 

[Ehwic]

Do you know why you'd use python over R? It seems like R is best for the bulk of analysis, while python may give you novel ways to organizing and differentiate between data sets.

Edit: typos... many typos

I'm not a comp sci major or anything, but here's what I found in terms of Python vs R for data analysis:



source: https://www.quora.com/Which-is-better-for-data-analysis-R-or-Python

Im not a biostatistician per se but I do have working knowledge with SAS, R, and Python.

SAS is the most different (or unique) in terms of environment and syntax out of all 3 packages.

R and Python syntax-wise is actually very similar with minor differences here and there. Their core concepts are mostly the same though. Where R excels is the statistical packages that it provides (not surprising as it was built that way.) I also like the visual display of R better than the other two programs.

Python, given that it's a programming language, lacks the more complex statistical power (however, this is slowly changing), but for data management, I found it to be more efficient and quicker for big data both in terms of analysis and management.

In my own experience, picking a package to work with is based on preference. I prefer SAS for all the data management and analysis as I have years of professional experiences with the program.

My second personal preference is Python because:

1) I like the syntax better than R.
2) I figured SAS can do all my statistical analysis so I dont need another statistical package.
2) It's a programming language so it allows me to automate a lot of my procedures/SAS codes.

Between Python and R, it really depends on your analysis. If you are dealing with complex data such as genomics then R is the better package. If it's simple analysis like regression models, ANOVAS, (simple) Mixed Models, Python would be sufficient.

But at the end, one should also consider their statistical knowledge. If you have a strong understanding of the theory behind stats then you can basically use either package and get the same result. Efficiency may vary but the end result is the same.

 

[Medic741]

I've been told that python is a must know for bioinformatics. How different is coding in R and python?

Haven't encountered a situation yet where I've needed something that R can't provide. My 'thing' in stats is metagenomics though, so pretty 'narrow' set of skills right there that R, QIIME, PiCRUST, and others pipelines cover pretty well.

You'll find that for many applications there are already pipelines available for your data workup, making learning python null for the majority of questions you'll be asking IMO

 

[cactusman]

R you sure this is the best language to use? I don't want to bash it, just asking to C if you're certain.

If you run on caffeine, I imagine the most important thing for productivity is just getting a cup of java.

 

[Lawpy]

R you sure this is the best language to use? I don't want to bash it, just asking to C if you're certain.

If you run on caffeine, I imagine the most important thing for productivity is just getting a cup of java.

 

[cactusman]

I can only see the photo when I go to quote your post. Maybe the Linux is broken?

 

[Endoxifen]

Try Swirl for R

This is excellent advice. It walks you through all the major functions in actual R. There is a JHU course on coursera or ed x about R as well if you are so inclined.

Thank you both for recommending this! I've been using it and it's an amazing program. R really isn't difficult at all and this makes it incredibly easy to learn.

 

[CyrusHabs7695]

Our end of the semester project for Biostats is learning a new package in R and presenting it to the class. Does anyone have any easy recommendations besides ggplot and the statistical tests (ANOVA, lm, glm etc.) Right now Im thinking of doing plot3D, but if anybody has any easy ones in mind thatd be cool

 

[Medic741]

Our end of the semester project for Biostats is learning a new package in R and presenting it to the class. Does anyone have any easy recommendations besides ggplot and the statistical tests (ANOVA, lm, glm etc.) Right now Im thinking of doing plot3D, but if anybody has any easy ones in mind thatd be cool

Not a package, bit time series regressions are super cool in R, your prof might be ok with this, also here's a great tutorial
http://www.statoek.wiso.uni-goettingen.de/veranstaltungen/zeitreihen/sommer03/ts_r_intro.pdf

A machine learning package, if you're feeling ambitious (not very hard)
https://cran.r-project.org/web/packages/dismo/vignettes/brt.pdf

 

[LePepe]

Our end of the semester project for Biostats is learning a new package in R and presenting it to the class. Does anyone have any easy recommendations besides ggplot and the statistical tests (ANOVA, lm, glm etc.) Right now Im thinking of doing plot3D, but if anybody has any easy ones in mind thatd be cool

I used the PK package for my lab. It was pretty easy to use. IDK how impressive it'll be though

 

[freak7]

10/10 LaTeX reference