Introdution to R

With special thanks to...

• Oliver Nakoinz
• Clemens Schmid
• Dirk Seidensticker
• Martin Hinz

• R is a language and environment for statistical computing and graphics
• Open Source Software: www.r-project.org
• extendible by packages (libraries)
• running with most operational systems
• interfaces to C, Python and Fortran etc.
• functional with object oriented extension
• vector based / dataframe
• developing and implementing own functions

• Data analysis is an essential tool of empirical research

• "learning programming is learning to think"

• forcing the problem in a logical and sequential structure

• 1976: S language by John Chambers at Bell Laboratories (AT&T)

• 1976: S language by John Chambers at Bell Laboratories (AT&T)

• 1993: Free implementation of R under the name R by Ross Ihaka and Robert Gentleman

• 1976: S language by John Chambers at Bell Laboratories (AT&T)
• 1993: Free implementation of R under the name R by Ross Ihaka and Robert Gentleman

• 1997: R becomes GNU project
• 2000: R1.0
• 2015: R Consortium: R Foundation, Microsoft, R Studio, Google etc.

Most statistical software provides convenient standartised test, as \(chi^2\) test etc. R has some very convienient advantages:

1. Asking a colleague to run R for you
2. Applying build in functions
3. Combining build in functions
4. Writing own algorithms
5. Writing own functions and objects
6. Developing efficient code
7. Writing own packages

1. Problem: Put your question.
2. Theory: Which theoretical frame is adequate for solving the problem?
3. Data: Which data can be used for solving the problem? Are they available as database etc.? How can you access the data?
4. Method: Which methods can solve the problem? Learn the methods!
5. Tool: Learn the tools to apply the methods.
6. Finally: Check, validate and interpret the results. Plausible? Significant?

• Scripts are an exact documentation of the work.
• Scripts allow reproducing the analysis.
• Scripts allow reusing commands, algorithms etc.
• Scripts allow sharing code.
• Scripts allow the automation of complex analyses.
• Scripts allow the automation of report generation; for instance, in R with the packages knitr or sweave.
• Scripts allow rerunning and testing code snippets.
• Scripts make it easy to create complex solutions.
• Scripts allow limiting the active knowledge of the researcher to the most important aspects.

• Each analysis has its own script.
• Each script has a header containing important information such as title, topic, purpose, author, date, etc.

 ##################################################################
 ## Didactic R-Script for Modelling Summer School
 ## ===============================================================
 ## Project: Modelling Summer School
 ## Author: O. Nakoinz, D. Knitter & F. Faupel
 ## Version: 01
 ## Date of last changes: 03.08.2016
 ## Data: some.data
 ## Author of data: author.data
 ## Purpose: didactic
 ## Content: 1. preparation, 2. data import, ...
 ## Description: The script include ...
 ## Licence data: -
 ## Licence Script: GPL (http://www.gnu.org/licenses/gpl-3.0.html)
 ##################################################################

• For each command, you should learn the usage only once.
  • Supplement the command with comments so that you can re-use the command for similar purposes without reading the manual again.
  • Copy, paste and adapt the commands from old scripts.

# This is a comment
a <- c(3,6,7,4,9,7,3,3,3)       # This is a vector
median(a)                         # This is the "median" function
## [1] 4

• Use standardised names and denotations to ease the re-use of code.

• Use detailed comments to understand your analysis years later.
  • It is useful to think that you should hand over the script to a colleague with limited knowledge to optimise comments and structure.
• Use versioning of the scripts by adding a version number to the file name (-v01.xxx) or a version control system.

• Use detailed comments to understand your analysis years later.
  • It is useful to think that you should hand over the script to a colleague with limited knowledge to optimise comments and structure.
• Use versioning of the scripts by adding a version number to the file name (-v01.xxx) or a version control system.
  • For example GitHub: https://github.com/

• Avoid special characters in file names, variable names, etc.
  • Subset Ä with Ae etc.
  • Moreover, it is a good idea to encode the file in UTF8.
• Use a convenient yet standardised folder structure for scripts, data, results and reports.

Use a convenient yet standardised folder structure for scripts, data, results and reports. During the Mosaic Summer School, we will work with the follwoing structure.

• J. Adler, R in a nutshell, Second edition. (Sebastopol 2012)
• J. M. Chambers, A Statistical Data Language. In: R. C. Milton, J. A. Nelder (Hrsg.), Statistical Computation. (New York 1969), 179-199.

• Scripts are an exact documentation of the work.
• Scripts allow reproducing the analysis.
• Scripts allow reusing commands, algorithms etc.
• Each analysis has its own script.
• Each script has a header containing important information such as title, topic, purpose, author, date, etc.
• Avoid special characters in file names, variable names, etc.
• Use a convenient yet standardised folder structure for scripts, data, results and reports.

# This is R Code 
5+3
5-3
5*3
5/3
sqrt(25)
sin(3.14)
3.14 * (180 / pi)

• Please start R Studio

• We will get used to R!

# This is R Code 
5+3
5-3
5*3
5/3
sqrt(25)
sin(3.14)
3.14 * (180 / pi)

# This is R Code
5+3
## [1] 8
5-3
## [1] 2
5*3
## [1] 15
5/3
## [1] 1.666667
sqrt(25)
## [1] 5
sin(3.14)
## [1] 0.001592653
3.14 * (180 / pi)
## [1] 179.9087