OTHER PLACES OF INTEREST
Danny Flamberg's Blog
Danny has been marketing for a while, and his articles and work reflect great understanding of data driven marketing.
Eric Peterson the Demystifier
Eric gets metrics, analytics, interactive, and the real world. His advice is worth taking...
Geeking with Greg
Greg Linden created Amazon's recommendation system, so imagine what can write about...
Ned Batchelder's Blog
Ned just finds and writes interesting things. I don't know how he does it.
R at LoyaltyMatrix
Jim Porzak tells of his real-life use of R for marketing analysis.
HOW DID YOU GET HERE?
First off, if you found this page via a web search or bookmark, you may be much happier in the R Section of this site to see the multiple articles about R, including this one, but also about Packages, Data manipulation, etc.
The home of all R GUIs is http://www.sciviews.org/_rgui/ which is accessible via the “R GUIs” link on the R home page. Start at the Overview link on the left nav. They include many Linux options which I will not delve into here at this time (I used to work at Microsoft, sorry bout that). Also linked are web front ends, SOAP and other RPC type calls of R servers, and DCOM/App links for Excel, Gnumeric and a few others.
This is a totally subjective list.
JGR is my current favorite. Running in Java, it includes completion, color syntax highlighting, integration of the iPlot interactive plotting packages, and nice people on the mailing list. I say start here, if for no other reason than its quick and easy. BTW, in the syntax editor, highlight your code and Ctrl-Enter to run it. Not documented, but that’s how its done.
update.packages()can be used to update the thing since they release new versions relatively frequently.
R-Commander is probably the most popular GUI these days. It does stay on the more “basic” side, but still is a good attempt to ease beginners into the world of R. And, after looking through the menus, there are some pretty handy shortcuts there, akin to SPSS Base offerings plus a few more. There are a bunch of plugins which extend it in different directions. Worth trying. Open Source.
SciViews-R is another popular gui. Currently in an alpha state, it could be a very powerful tool, though it feels more like an IDE than an analytic home. It is getting constant revision, and lots of attention on the mailing lists. One nice feature: A “report editor” with full HTML editing capabilities. In addition, SciViews can integrate with R-Commander and include its menus and features as well. All of this is open source.
If you just want an editor, the IDE/Script Editors page lists many of them, including Jedit. One interesting quickie is Tinn-R, basically a Notepad with some R twists and tweaks. I tend to default to Jedit, but that’s because I use it in so many other places. In addition, JGR has color syntax and code completion, so I have been using it more often as an editor as well. As I said above, however, if you want a light code editor, Tinn-R is pretty good.
Up and Comers:
RKward is an open-source GUI which looks pretty good… but is KDE and therefore strongest on Linux. Thanks to the porting of some libraries, it can run on Windows (barely), but it really needs help. Linux can handle memory better than Windows, so you might find yourself forced to use a Linux box for some analyses. You’ll want to use RKward. Once it gets better on Windows, it will be a powerhouse. See docs at http://sourceforge.net/apps/mediawiki/rkward/index.php?title=Main_Page
PMG is the “poor man’s gui”. I haven’t tried it yet, requires GTK libraries (Linux will already have them, Windows will need them installed; site tells you how). It’s getting some updates, but slowly, so YMMV.
Rattle which Togaware (an Australian Data Miner) says “(the R Analytical Tool To Learn Easily) provides a simple and logical interface for quick and easy data mining.” Don’t know much about it; it too requires the GTK libraries which are pretty easy to install. Gets lots of love on the mailing lists, partially from it’s ease, and partially that it brings in lots of data mining components often ignored in more “statistics” focused packages. Lots of great content on this site about the process of actually data mining.
SimpleR GUI is a good start for a GUI. Simple-R Handbook (PDF) shows how to use it. Some very clever features and graphics. However, it hasn’t been updated in a while, so it may not work as well as you wish.
If you have money, REvolution R Enterprise is an Eclipse based IDE which is pretty nice. Get your company to pay for it. Also, Eclipse can be a bit of a pig, so a beefier box is helpful.
Couldn’t care less about:
ESS and the Emacs bindings: as a non-emacs person (go Jedit!), I find little value in using a fat and overblown editor to run an already complicated system. Talking to emacs/vim folks (they both have problems) is similar to Name that Tune in reverse: I can do that in 10 keystrokes! I can do that in 12 keystrokes! Anyway… I am really just jealous because I’ve tried a few times and still can’t learn Emacs or VI, so I derogate them to make myself feel better. Anyway. If you already love Emacs, ESS is the way to go.
Brogdar is a commercial app which sits on top of R. However, it costs around $350, and though it has some interesting features, I can’t recommend it at this point (for $350, you can start to buy entire stats packages with support, so best of luck to them).
OpenI is an overall reporting and BI system, but it has a link to use R. This is by the astounding guys at LoyaltyMatrix [now Responsys]. See more of their coolness at R at LoyaltyMatrix
BEE is another suite of tools supporting Business Intelligence project implementation including ETL tool and OLAP server and a thin client, with R as the “analytic” engine. Not updated since 2005, which is too bad.
DecisionStudio is a “desktop BI platform”. Basically integrates open source offerings including MySql, DBDesigner, R and Tinn-R, iReport and JasperReports, all linked with Python. Last updated in March, 2006, so may be dead.
Little Seeds, starting to grow…
Biocep-R, Statistical Analysis Tools for the Cloud Computing Age includes not only a Java front end, but a whole framework for distributed computing. The “Virtual R Workbench” looks pretty impressive, but unclear how much of the other server stuff you need.
R AnalyticFlow is like the visual front end from Clementine or SAS Ent Miner. Still early, but could be the front end to keep if you like these flowchart approaches.
RedR is a cool dataflow approach, open source, still very young.
Visualization of Data
Besides a gui to help you use R, there is a also a growing field of graphical depiction of data to help with analysis. This can be as simple as a scatterplot, or as sophisticated as a 3d chart allowing you to select points and see their relationships highlighted on multiple other charts simultaneously.
One of the more popular tools for this in the open source world is the GGOBI system, and R has a pretty good linkup called rggobi. Like R, it’s a whole system, so you won’t get all of its power on the first day, but it’s worth playing with to see just how easy it can be.
(Thanks to Andy Edmonds on the Web Analytics Group for reminding me that I left rggobi out.)
This isn’t exactly the same, but if you already use a package, you might find that it wraps around R. So, for example, SPSS has the R Extension available from their installer CDs in recent versions, or at SPSS Developer Central. If you use SAS/IML (which is pretty hard core, it’s only if you mess with your own matrices), you can get to R via their SAS/IML to R Interface. Other tools are also starting to wrap around R.
Rpad is an interesting mix of web based tool with more of an app feel. As they say on their page, “Rpad is an interactive, web-based analysis program. Rpad pages are interactive workbook-type sheets. Rpad is an analysis package, a web-page designer, and a gui designer all wrapped in one. The user doesn’t have to install anything—everything’s done through a browser.” Could become something cool in the future… though it appears to be gone now.
* * *