What R packages do you find most useful in your daily work?

Question

Duplicate thread: I just installed the latest version of R. What packages should I obtain?

What are the R packages you couldn't imagine your daily work with data? Please list both general and specific tools.

UPDATE: As for 24.10.10 ggplot2 seems to be the winer with 7 votes.

Other packages mentioned more than one are:

plyr - 4
RODBC, RMySQL - 4
sqldf - 3
lattice - 2
zoo - 2
Hmisc/rms - 2
Rcurl - 2
XML - 2

Thanks all for your answers!

Very subjective question: this question cannot be answered, and is not suitable for a QA site. — Egon Willighagen, Jul 19 '10 at 19:58
Should probably be community wiki; useful question here but doesn't have definitive answer. — Shane, Jul 19 '10 at 20:05
@Shane: good point. moved. @ Egon: subjective indeed. but if the answers come from knowledgeable people i don't mind dose of subjectivity. i've started learning R quite recently and have couple of dozens installed to explore, however i notice that there are tools that I use much more often irrespectively of the task at hand. — radek, Jul 19 '10 at 20:06
It would be interesting if StackExchange could support some method of linking community wiki posts across sites. Because I will bet this question has been asked on Stackoverflow and I also think that Statistical Analysis may attract some people that wouldn't usually visit SO. — Sharpie, Jul 19 '10 at 20:19
@Sharpie: there have been several interesting SO posts like http://stackoverflow.com/questions/1295955/what-is-the-most-useful-r-trick or http://stackoverflow.com/questions/1535021/whats-the-biggest-r-gotcha-youve-run-across however they are not focused on packages. and i agree, linkage of community wiki could be really useful. — radek, Jul 19 '10 at 20:37
I disagree that it's for SO, especially considering the discussions on meta that supporting tools for statistical analysis (including software) is on-topic.... — Thomas Owens, Aug 14 '10 at 12:25
I vote to close because "his question is likely to be answered with opinions rather than facts and citations.". — Bernhard, Oct 19 '20 at 08:33

score 23 · Answer 1 · answered Jul 19 '10 at 19:52

23

I use plyr and ggplot2 the most on a daily basis.

I also rely heavily on time series packages; most especially, the zoo package.

answered Jul 19 '10 at 19:52

Shane

12,461

score 12 · Answer 2 · answered Aug 14 '10 at 13:06

12

In a narrow sense, R Core has a recommendation: the "recommended" packages.

Everything else depends on your data analysis tasks at hand, and I'd recommend the Task Views at CRAN.

answered Aug 14 '10 at 13:06

Dirk Eddelbuettel

8,872

1

+1 Package demands heavily depend on each individual and I think Task Views are the way to start anyone. – Roman Luštrik May 11 '11 at 11:32

Stephen Turner · Answer 3 · 2010-07-20T15:28:42.347

ggplot2 - hands down best visualization for R.

RMySQL/RSQLite/RODBC - for connecting to a databases

sqldf - manipulate data.frames with SQL queries

Hmisc/rms - packages from Frank Harrell containing convenient miscellaneous functions and nice functions for regression analyses.

GenABEL - nice package for genome-wide association studies

Rcmdr - a decent GUI for R if you need one.

Also check out CRANtastic - this link has a list of the most popular R packages. Many on the top of the list have already been ment

score 8 · Answer 4 · answered Jul 27 '10 at 04:27

8

data.table is my favorite now! Very look forward to the new version with the more wishlist implemented.

answered Jul 27 '10 at 04:27

score 8 · Answer 5 · answered Jul 19 '10 at 19:58

8

I use the xtable package. The xtable package turns tables produced by R (in particular, the tables displaying the anova results) into LaTeX tables, to be included in an article.

answered Jul 19 '10 at 19:58

Mehper C. Palavuzlar

1,164

score 8 · Answer 6 · answered Jul 19 '10 at 20:25

8

multicore is quite nice for tool for making faster scripts faster.
cacheSweave saves a lot of time when using Sweave.

answered Jul 19 '10 at 20:25

score 6 · Answer 7 · answered Jul 30 '10 at 15:43

6

Sweave lets you embed R code in a LaTeX document. The results of executing the code, and optionally the source code, become part of the final document.

So instead of, for example, pasting an image produced by R into a LaTeX file, you can paste the R code into the file and keep everything in one place.

answered Jul 30 '10 at 15:43

John D. Cook

4,092

4

Just a hint for all those who want to get started on reproducible research with R. I would advice you to have a look at the newish package knitr instead of Sweave. It's basically Sweave on steroids. It is as easy, if not easier, to learn and far more flexible. – Christoph_J Feb 18 '13 at 21:05

score 6 · Answer 8 · answered Aug 14 '10 at 01:46

I imagine graphics and data manipulation are two things that are useful no matter what you are doing. Thus, I'd recommend:

ggplot2 (great graphics)
lattice (great graphics)
plyr (useful for data manipulation)
Hmisc (good for descriptive statistics and much more)

Tony Breyal · Answer 9 · 2010-07-28T10:02:05.600

For me personally, I use the following three packages the most, all available from the awesome Omega Project for Statistical Computing (I do not claim to be an expert, but for my purposes they are very easy to use):

RCurl: It has lots of options which allows access to websites that the default functions in base R would have difficulty with I think it's fair to say. It is an R-interface to the libcurl library, which has the added benefit of a whole community outside of R developing it. Also available on CRAN.
XML: It is very forgiving of parsing malformed XML/HTML. It is an R-interface to the libxml2 library and again has the added benefit of a whole community outside of R developing it Also available on CRAN.
RJSONIO: It allows one to parse the text returned from a json call and organise it into a list structure for further analysis.The competitor to this package is rjson but this one has the advantage of being vectorised, readily extensible through S3/S4, fast and scalable to large data.

score 6 · Answer 10 · answered Jul 27 '10 at 09:00

6

Packages I often use are raster, sp, spatstat, vegan and splancs. I sometimes use ggplot2, tcltk and lattice.

answered Jul 27 '10 at 09:00

Roman Luštrik

3,718

score 4 · Answer 11 · answered May 08 '11 at 02:26

4

If you are doing any kind of predictive modeling, caret is a godsend. Especially combined with the multicore package, some pretty amazing things are possible.

answered May 08 '11 at 02:26

Zach

23,766

score 4 · Answer 12 · answered Aug 14 '10 at 14:06

4

You can get user reviews of packages on crantastic

answered Aug 14 '10 at 14:06

Karsten W.

697
6
25

score 4 · Answer 13 · answered Aug 14 '10 at 21:52

4

I would suggest using some of the packages provided by revolution R. In particular, I quite like the:

multicore package for parallel computing using shared memory processors
there optimized packages for matrices

answered Aug 14 '10 at 21:52

csgillespie

13,029

score 4 · Answer 14 · edited Aug 27 '12 at 17:52

4

Day-to-day the most useful package must be "foreign" which has functions for reading and writing data for other statistical packages e.g. Stata, SPSS, Minitab, SAS, etc. Working in a field where R is not that commonplace means that this is a very important package.

edited Aug 27 '12 at 17:52

whuber

322,774

answered Jul 20 '10 at 11:59

Graham Cookson

8,061

score 4 · Answer 15 · answered Jul 27 '10 at 07:24

4

zoo and xts are a must in my work!

answered Jul 27 '10 at 07:24

Dr G

1,214

score 4 · Answer 16 · answered Jul 27 '10 at 09:51

4

I find lattice along with the companion book "Lattice: Multivariate Data Visualization with R" by Deepayan Sarkar invaluable.

answered Jul 27 '10 at 09:51

NPE

5,581
6
37
45

score 3 · Answer 17 · answered May 10 '11 at 23:50

3

I use lattice, ggplot2, lubridate, reshape, boot, e1071, car, forecast, and zoo a lot.

answered May 10 '11 at 23:50

Wayne

21,174

score 3 · Answer 18 · answered Aug 14 '10 at 04:58

This is definitely a question that doesn't have "an answer". It is completely dependent on what you want to do. That aside, I'll share the packages that I install as a standard with an R update...

install.packages(c("car","gregmisc","xtable","Design","Hmisc","psych",
                        "CCA", "fda", "zoo", "fields",
                      "catspec","sem","multilevel","Deducer","RQDA"))

and leave it to you to investigate those packages and see if they are valuable to you.

score 3 · Answer 19 · edited May 10 '11 at 22:15

3

If you are working with Latex, I recommend TikZ Device for outputting nice, Latex-formatted (like PSTricks) graphics. The output you get is text-based Latex code, which can be embedded with include(filename) into any figure environment.

Pros:

Same font in graphics as in your text
Professional look

Cons:

Takes longer to compile than PNG or PDF
for very complex R graphics, there are could be some display errors

https://github.com/Sharpie/RTikZDevice - Project, Packages available from CRAN and R-Forge

edited May 10 '11 at 22:15

Sharpie

4,374

answered Aug 14 '10 at 08:51

dermatthias

139

2

No disrespect, but that strikes me as entirely subjective. I have used latex for since the late 80s and R since the 90s yet I see no good reason to move away from embedding eps files. Everybody's mileage will differ here, and Thomas should maybe look at Task Views for particular problem domains. – Dirk Eddelbuettel Aug 14 '10 at 13:04
@Dirk: Good typography should not draw attention to its self and getting different text from LaTeX and R graphics is very visible (EPS output on the left, Sweave source code). Admittingly the example is a bit biased as no effort is made to customize the font used by eps. The point is that tikz should match the LaTeX document font, font size, letterspacing, kerning, ligaturing, etc. by default. If it doesn't, I consider it a bug. – Sharpie May 10 '11 at 21:54

score 3 · Answer 20 · answered Sep 01 '10 at 11:31

3

You can also take a look at Task views on CRAN and see if something suit your needs. I agree with @Jeromy for these must-have packages (for data manipulation and plotting).

answered Sep 01 '10 at 11:31

chl

53,725

score 3 · Answer 21 · answered Aug 27 '12 at 05:03

I could not live without:

lattice for graphics
xlsx or XLConnect for reading Excel files
rtf to create reports in rtf format (I would prefer Sword or R2wd but I cannot install statconn at work; I will surely try odfWeave soon.)
nlme and lme4 for mixed models
ff for working with large arrays

score 3 · Answer 22 · answered Oct 19 '20 at 07:15

3

I can recommend the new shiny based packages to everyone, it makes data visualisation and inspection interactive and thus easier than writing code in R espacially in the beginning.

A good example would be ggplotgui

answered Oct 19 '20 at 07:15

SeppSeppenbauer

43

score 3 · Answer 23 · answered Jul 27 '10 at 08:42

3

I use

car, doBy, Epi, ggplot2, gregmisc (gdata, gmodels, gplots, gtools), Hmisc, plyr, RCurl, RDCOMClient, reshape, RODBC, TeachingDemos, XML.

a lot.

answered Jul 27 '10 at 08:42

score 2 · Answer 24 · answered May 08 '11 at 11:53

2

lattice, car, MASS, foreign, party.

answered May 08 '11 at 11:53

Peter Flom

119,535
36
175
383

score 2 · Answer 25 · answered May 11 '11 at 12:33

2

For me I am using kernlab for Kernel-based Machine Learning Lab and e1071 for SVM and ggplot2 for graphics

answered May 11 '11 at 12:33

404Dreamer_ML

398

score 2 · Answer 26 · answered Sep 21 '14 at 10:41

I am a big fan of RCPP when I need a rapid for loop or to perform non very R compliant treatments. It is very well implemented in the R eco system, can receive Matrix / sparse Matrix without conversion as arguments in a function.

C++ syntax is easy when you are doing simple stuff (which is often my case).

Really, you don't need to be a package maker to need this awesome lib.

Did I say C++ is very rapid?

score 2 · Answer 27 · answered Jun 17 '11 at 00:57

2

I use ggplot2, vegan and reshape quite often.

answered Jun 17 '11 at 00:57

EDi

1,171

score 2 · Answer 28 · answered Jun 17 '11 at 02:29

2

I like roxygen for its Curry() function.

answered Jun 17 '11 at 02:29

charles.y.zheng

7,936

score 2 · Answer 29 · answered Aug 20 '15 at 09:00

The doParallel and foreach packages have made my life so much easier by allowing me to parallelize my code and run it on a compute-optimized instance on Amazon EC2! I use them very often. But that would not have been possible without the RStudio AMIs released by Louis Aslett. Finally, I have to mention the stringr package which really makes working with strings a walk in the park. Use it in every text mining application. And I also use knitr very frequently to produce high quality reports of my work. Many thanks for this amazing package Yihui Xie!

score 2 · Answer 30 · edited Mar 06 '14 at 00:03

2

Jeromy mentioned my first pick: Lattice.

I also have found the doBy package and its summaryBy function to be insanely useful. They extend aggregate with a formula syntax that lets you aggregate multiple functions simultaneously in non-trivial ways. Great if you want, say, mean, std. dev., and length.

edited Mar 06 '14 at 00:03

Andre Silva

3,080

answered Aug 15 '10 at 00:31

Michael Ekstrand

103

score 2 · Answer 31 · answered Aug 27 '12 at 07:28

2

RColorBrewer has not been mentioned here, I use it often for plotting if I need color schemes

answered Aug 27 '12 at 07:28

Arne Jonas Warnke

3,513

score 2 · Answer 32 · edited Aug 09 '10 at 08:15

2

We mostly use:

ggplot - for charts
stats
e1071 - for SVMs

edited Aug 09 '10 at 08:15

Peter Mortensen

343
3
10

answered Jul 27 '10 at 16:37

mohans

283
3
5

You may also want to check out kernlab and caret for SVMs. They're interesting (thought not necessarily better) alternatives. – Zach May 08 '11 at 02:27

score 2 · Answer 33 · edited Aug 09 '10 at 08:13

2

I work with both R and MATLAB and I use R.matlab a lot to transfer data between the two.

edited Aug 09 '10 at 08:13

Peter Mortensen

343
3
10

answered Jul 27 '10 at 18:33

Shrey

146

score 2 · Answer 34 · answered Jul 28 '10 at 13:59

2

RODBC for accessing data from databases, sqldf for performing simple SQL queries on dataframes (although I am forcing myself to use native R commands), and ggplot2 and plyr

answered Jul 28 '10 at 13:59

wahalulu

171

score 1 · Accepted Answer · answered Jun 23 '14 at 08:58

1

Please see link: TOP 100 R PACKAGES FOR 2013 (JAN-MAY) http://www.r-statistics.com/2013/06/top-100-r-packages-for-2013-jan-may/

answered Jun 23 '14 at 08:58

adam.888

123

score 1 · Answer 36 · answered May 11 '11 at 10:03

1

Some packages are very useful in R.

I will just recommand kernlab for Kernel-based Machine Learning Lab and e1071 for SVM and ggplot2 for graphics

answered May 11 '11 at 10:03

404Dreamer_ML

398

score 1 · Answer 37 · answered Feb 18 '13 at 17:18

1

I use ggplot2, reshape, lattice, knitr more often.

answered Feb 18 '13 at 17:18

Jdbaba

101

What R packages do you find most useful in your daily work?

37 Answers37

Linked

Related