An Introduction to R. Subhajit Dutta Stat-Math Unit. Indian Statistical Institute, Kolkata October 17, 2012

Transcription

1 An Introduction to R Subhajit Dutta Stat-Math Unit Indian Statistical Institute, Kolkata October 17, 2012

2 Why R? It is FREE!! Basic as well as specialized data analysis technique at your fingertips. Highly competitive with existing expensive statistical packages (like Matlab, SAS, SPSS). Very small installer file ( 30 MB). Easy to write codes. Draw on the talents of data scientists worldwide.

3 Background S was developed by John Chambers and others at Bell Labs, 1976 as an internal statistical analysis environment. In 1993, Bell Labs gave Stat Sci (now Insightful Corp.) an exclusive license to develop, and sell the S language. R was created by Ross Ihaka and Robert Gentleman at the University of Auckland, New Zealand, in 1991.

4 Background S was developed by John Chambers and others at Bell Labs, 1976 as an internal statistical analysis environment. In 1993, Bell Labs gave Stat Sci (now Insightful Corp.) an exclusive license to develop, and sell the S language. R was created by Ross Ihaka and Robert Gentleman at the University of Auckland, New Zealand, in In 1997, the R-Core Group was formed. The first R version is released in 2000.

5 Background S was developed by John Chambers and others at Bell Labs, 1976 as an internal statistical analysis environment. In 1993, Bell Labs gave Stat Sci (now Insightful Corp.) an exclusive license to develop, and sell the S language. R was created by Ross Ihaka and Robert Gentleman at the University of Auckland, New Zealand, in In 1997, the R-Core Group was formed. The first R version is released in Chambers is now a member of the R-Core Group.

6 Background S was developed by John Chambers and others at Bell Labs, 1976 as an internal statistical analysis environment. In 1993, Bell Labs gave Stat Sci (now Insightful Corp.) an exclusive license to develop, and sell the S language. R was created by Ross Ihaka and Robert Gentleman at the University of Auckland, New Zealand, in In 1997, the R-Core Group was formed. The first R version is released in Chambers is now a member of the R-Core Group. R is named partly after the first names of the first two R authors, and partly as a play on the name of S.

7 Developer of S : John Chambers

8 The Developers of R : Ross Ihaka and Robert Gentleman

9 Design of R The R system is divided into 2 conceptual parts :

10 Design of R The R system is divided into 2 conceptual parts : The base R system.

11 Design of R The R system is divided into 2 conceptual parts : The base R system. User created packages : Allows specialized statistical techniques.

12 Design of R The R system is divided into 2 conceptual parts : The base R system. User created packages : Allows specialized statistical techniques. R functionality is divided into a number of packages, more than 4000 (as of August 2012) are available.

13 Design of R The R system is divided into 2 conceptual parts : The base R system. User created packages : Allows specialized statistical techniques. R functionality is divided into a number of packages, more than 4000 (as of August 2012) are available. Download from Comprehensive R Archive Network (CRAN), which contains mirrors throughout the world.

14 Design of R The R system is divided into 2 conceptual parts : The base R system. User created packages : Allows specialized statistical techniques. R functionality is divided into a number of packages, more than 4000 (as of August 2012) are available. Download from Comprehensive R Archive Network (CRAN), which contains mirrors throughout the world. India hosts a mirror at IIT, Madras.

15 Design of R The R system is divided into 2 conceptual parts : The base R system. User created packages : Allows specialized statistical techniques. R functionality is divided into a number of packages, more than 4000 (as of August 2012) are available. Download from Comprehensive R Archive Network (CRAN), which contains mirrors throughout the world. India hosts a mirror at IIT, Madras. Dr. Deepayan Sarkar, ISID (tomorrow s speaker) is among the core developers for R.

16 Areas of application Applications are listed at :

17 Areas of application Applications are listed at : Statistical Pattern Recognition, Spatial and Bayesian Statistics.

18 Areas of application Applications are listed at : Statistical Pattern Recognition, Spatial and Bayesian Statistics. Finance, Genetics, Machine Learning, Medical Imaging and Social Sciences.

19 Areas of application Applications are listed at : Statistical Pattern Recognition, Spatial and Bayesian Statistics. Finance, Genetics, Machine Learning, Medical Imaging and Social Sciences. The Bioconductor (hosted by Fred Hutchinson Cancer Research Center, USA) project provides 460 packages. Microarrays, High Throughput Assays, Sequence Data and Annotation.

20 Areas of application Applications are listed at : Statistical Pattern Recognition, Spatial and Bayesian Statistics. Finance, Genetics, Machine Learning, Medical Imaging and Social Sciences. The Bioconductor (hosted by Fred Hutchinson Cancer Research Center, USA) project provides 460 packages. Microarrays, High Throughput Assays, Sequence Data and Annotation. Google uses R to make online advertising more effective. Statistics : The Secret Weapon of Successful Web Giants (JSM, 2011).

21 Powerful and Excellent Visualizations

22 Powerful and Excellent Visualizations

23 Powerful and Excellent Visualizations

24 Advantages of using R A lot of packages are uploaded by the authors/their students.

25 Advantages of using R A lot of packages are uploaded by the authors/their students. It contains advanced statistical routines not yet available in other softwares.

26 Advantages of using R A lot of packages are uploaded by the authors/their students. It contains advanced statistical routines not yet available in other softwares. Workspace can be saved, and shifted to another some other computer.

27 Advantages of using R A lot of packages are uploaded by the authors/their students. It contains advanced statistical routines not yet available in other softwares. Workspace can be saved, and shifted to another some other computer. Lines of script are significantly less, and hence easier for debugging.

28 Advantages of using R A lot of packages are uploaded by the authors/their students. It contains advanced statistical routines not yet available in other softwares. Workspace can be saved, and shifted to another some other computer. Lines of script are significantly less, and hence easier for debugging. Codes in C can be interfaced with R, and vice versa.

29 Advantages of using R A lot of packages are uploaded by the authors/their students. It contains advanced statistical routines not yet available in other softwares. Workspace can be saved, and shifted to another some other computer. Lines of script are significantly less, and hence easier for debugging. Codes in C can be interfaced with R, and vice versa. R respects C, it retains the speed of the C code when run in an R environment.

30 Advantages of using R A lot of packages are uploaded by the authors/their students. It contains advanced statistical routines not yet available in other softwares. Workspace can be saved, and shifted to another some other computer. Lines of script are significantly less, and hence easier for debugging. Codes in C can be interfaced with R, and vice versa. R respects C, it retains the speed of the C code when run in an R environment. R code/data written by you can be shared with the rest of the statistics community as a package.

31 Drawbacks, and comments from fellow users Cannot handle huge data sets.

32 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data.

33 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners.

34 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners. Slow

35 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners. Slow Specific commands can improve speed a lot.

36 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners. Slow Specific commands can improve speed a lot. Comments

37 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners. Slow Specific commands can improve speed a lot. Comments Support : send an to "rhelp@r-project.org" and you will get very good/fast response.

38 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners. Slow Specific commands can improve speed a lot. Comments Support : send an to "rhelp@r-project.org" and you will get very good/fast response. Anything related to Math/Stat/Finance, you will get an existing package/function.

39 Drawbacks, and comments from fellow users Cannot handle huge data sets. Usual R can easily handle 400 MB data. With the library bigmemory, there is no limit on data size. But, using it is not as easy for beginners. Slow Specific commands can improve speed a lot. Comments Support : send an to "rhelp@r-project.org" and you will get very good/fast response. Anything related to Math/Stat/Finance, you will get an existing package/function. You have the scope to explore.

40 R installation

41 Data Types, Arrays, Loops (avoid!!).

42 Matrix Manipulations.

43 Sorting and Searching.

44 Graph Plotting (2D and 3D).

45 Probability Distributions.

46 File Reading and Writing.

47 Integrating C with R / R with C.

48 The best thing about R is that it was developed by statisticians. The worst thing about R is that...

49 The best thing about R is that it was developed by statisticians. The worst thing about R is that... it was developed by statisticians. Bo Cowgill, Google.

50 The best thing about R is that it was developed by statisticians. The worst thing about R is that... it was developed by statisticians. Bo Cowgill, Google. R you ready for R? Thanks to Palash Da, Buddha, Minerva and of course, Google.