Info for R users (Links, Manuals, Books, etc.)

#81
For those R beginners who just want to graph right away.. This is a great link that plots your data with ggplot2 and gives you the code to do it yourself..

http://yeroon.net/ggplot2/

Just upload your data from a csv or excel or google data, right click to define variables and layers, and see your plot.. Many options are offered (though not the full capabilities of the grammar of graphics in R) and you can even export your plot as a pdf or svg.

All thanks to Jeroen C.L. Ooms and UCLA

info can be found here: http://www.stat.ucla.edu/~jeroen/

other web apps offered are an IRT tool and a LME4 tool :D
 

bryangoodrich

Probably A Mammal
#82
You'll find many of us hold conflicting opinions about turning to ggplot2. We all acknowledge it is a great utility, but a few of us (me included) hate the idea of saying "hey R n00b that doesn't know how to plot, go use this to make up for your ignorance." You need to learn how to walk before you run, and it is too easy for someone not to appreciate how to manage their data and understand why ggplot2 works the way it does. With that said, I'm also for making R and its utilities accessible for those that aren't going to be R programmers. That's the issue, though. Some people want to use R to achieve an end without knowing the ins and outs. Others want to be those "experts." I think it is best to defer to the experts to do the work. But I'm biased. It gives me career options! hahaha
 

bryangoodrich

Probably A Mammal
#83
With that said, I like the interface. That's the sort of interface that should be built right into RStudio to let people use a nice GUI for their R usage and be able to conveniently take advantage of the ggplot grammar of graphics to produce nice graphs.
 

trinker

ggplot2orBust
#84
I tend to disagree with Bryan (though to be fair he's presented a generally balanced though not unbiased statement(s)). I think that he's correct for people interested in data mining but for the researcher who wants to understand their data better without learning a ton of stuff this can have its place. You have x number of hours and have to choose where to spend your time learning as a researcher. For many they just can't afford to learn deeply in visualization/programming (though I may argue some time up front may save them a bunch of time later). This argument of decent tools that do a guided job without the researcher fully understanding the ins and outs is true for SPSS as well. For some people SPSS is a great tool (though like ggplot2 it can be misused). I actually believe that ggplot is harder to learn than base but there's way more functions/arguments (par is a loaded burrito) to learn with base. Bryan actually presents a fairly balanced view, however, I'd point out that sometimes the experts don't understand the content or there's not money in the budget for an expert. I think that the user has to put some effort into being informed but these types of tools can be very helpful for understanding data. Many researchers don't care that much about the awesomeness of the visual or the act of creating it, they are more focused on what that data is saying.
 
#85
I agree, but we all know what a painstaking process it is to get started with R.. Many give up before it even starts to click.. I see these web apps as learning tools. And in no way can a casual R user replace an expert statistician. I searched, there is no package that can interpret data and give meaningful conclusions about that data. That takes an in-depth knowledge and keen understanding of the methods of data analysis, which is a huge accomplishment in itself. This is just an option for a quick and easy to view one's data.. Without the added hours learning R syntax

I like the interface as well.. but many of the plot options were left out.

Edited... Didn't mean to sound mean.
 
Last edited:
#89
Hello all

Just a quick question - is this thread an OK place to post about my new book on R? I know that someone turning up on a forum and making a commercial announcement as a first post can sometimes be seen as rude, so if that would be the case here let me know and I'll keep quiet.

Cheers

Rob
 
#90
Since no-one's replied, can I just say that I have just published a book called "Introductory R: A Beginner's Guide to Data Visualisation and Analysis using R". Currently available as an ebook on Amazon and Kobo, Apple and Google Play coming soon.

It's aimed at people who know a little bit of stats but have no experience with coding or command line interfaces. The examples are from the life sciences but it should be accessible to people from other disciplines.

I've also put up a page of R resources for beginners which some people might find helpful.

Any comments, questions etc. let me know.

Rob K
 

bugman

Super Moderator
#96
TE, that link is pretty cool and I found it pretty useful.

one thing though -

The author suggests that using

attach()

is a bad idea. I have always used it, I just wonder why it would be considered a bad idea?
 

trinker

ggplot2orBust
#97
@Bugman It could cause you to write over information and not realize it. It also muddies up your workspace. Generally I've seen it done in books for ease of explanation bit in reality people tend to use an explicit data argument or the with function. I don't know how much it matters though if it works for your workflow.
 

TheEcologist

Global Moderator
@Bugman It could cause you to write over information and not realize it. It also muddies up your workspace. Generally I've seen it done in books for ease of explanation bit in reality people tend to use an explicit data argument or the with function. I don't know how much it matters though if it works for your workflow.
In cases where you are working with a single simple dataset (SPSS like), and you don't want to add columns or convert values in the data, there is no reason not to use it. Once things get more complicated than this.. don't.

But who am I? I'm not the great Tyler Rinker.

;-)