Introduction to R - Urfist · Introduction Data types and structures Control structures Data Handling Basic Analysis Introduction to R Ur st Moritz Mul ler, Ma^ tre de Conf erence

Introduction Data types and structures Control structures Data Handling Basic Analysis

Introduction to RUrfist

Moritz Muller, Maıtre de Conference at FSEG, Universit’e deStrasbourg

05.-06.12.2019

1 / 109


Introduction

Data types and structures

Control structures

Data Handling

Basic Analysis

2 / 109


OverviewIntroduction

What is R?First stepsExamplesData typesVectorMatrixListsData.framesConditional expressionsLoopsRead/WriteData cleaning/formattingData Merge/SelectionPlottingStatisticsBeyond this course

3 / 109


References

Shorted list

• Venables, W. N. and Ripley, B. D. (2002) Modern Applied Statisticswith S. Fourth Edition. Springer, New York. ISBN 0-387-95457-0(downloadable)

• Uwe Ligges (2009): Programmieren mit R, 3.,Uberarb. u.erweiterte Auflage, Springer Verlag, Heidelberg.

• Venables, W.N., Smith, D.M., and the R core team (2015) AnIntroduction to R. Notes on R: A Programming Environment forData Analysis and Graphics. Version 3.2.0 (2015-04-16). Downloadhttp://cran.r-project.org/doc/manuals/R-intro.pdf

• Daalgard, P (2008) Introductory Statistics with R. Second Edition.Springer Science+Business Media LLC, New York. Downloadhttp://www.academia.dk/BiologiskAntropologi/Epidemiologi/PDF/Introductory Statistics with R 2nd ed.pdf.

4 / 109


Why (not) use R?

5 / 109


What is R?

What is R?Object oriented, interpreted, language and programmingenvironment for Data analysis and graphics.

• huge collection of tools for statistics and data analysis

• a language for expressing statistical models and tools

• graphical facilities

• effective object-oriented programming language that can beextended

6 / 109


Short History of R

Short history

1976 S language developed by Bell Labs at AT&T

1988 S-PLUS, commercial version of S (Statistical Sciences Inc.)

1995 R, open source (GPL) version of S (by Ross Ihaka and RobertGentleman)

1997 R Development Core Team (in short: R Core Team)

1998 Comprehensive R Archive Network (CRAN) founded

2000 R-1.0.0 first version compatible with S

2001 R News journal appears

2004 R version R-2.0.0 (S4 methods in package methods)

2015 R consortium (industry support)

7 / 109


Short History of R - submitted packages over time1

1Gergely Daroczi, source: https://gist.github.com/daroczig8 / 109


Why (not) use R?

Interpreted language

• programming on the fly

• flexible handling

• slower than a compiled language such as C, C++ (manybuilt-in functions are in C)

9 / 109


Why (not) use R?

Open Source

• No black box (in principle)

• New methods early available

• MCMC with Stan (https://github.com/stan-dev/rstan/wiki/RStan-Getting-Started),

• Webscraping with rSelenium(https://www.seleniumhq.org/projects/webdriver/),

• Deep learning with Keras(https://www.tensorflow.org/guide/keras),

• Natural language processing with WordNet(https://wordnet.princeton.edu/), and many more.

• Support from the community

10 / 109


Why (not) use R?

Open Source & Programming needed

• No central authority

• No GUI (Graphical User Interface)

• e.g. offered by SPSS, SAS

Do you want to travel all-inclusive or as back-packer?

11 / 109


Why (not) use R?

Open Source & Programming needed

• No central authority

• No GUI (Graphical User Interface)

• e.g. offered by SPSS, SAS

Do you want to travel all-inclusive or as back-packer?

11 / 109


Today’s topicsIntroduction

What is R?First stepsExamples

Data types and structuresData typesVectorMatrixListsData.frames

Control structuresConditional expressionsLoops

Data HandlingRead/WriteData cleaning/formattingData Merge/Selection

Basic AnalysisPlottingStatisticsBeyond this course

12 / 109




13 / 109


Installation on your own device:

Install R

• go to http://www.r-project.org/

• Download R for Windows / Mac / Linux

• Install ‘Base distribution’ with default settings

Install RStudio

• go to http://www.rstudio.org/

• Download and install with default

14 / 109


Start R Studio

Explain R vs. ‘R Studio’. Small tour through RStudio.

15 / 109


R - Calculator

Math expressions

> 1 + 2 * 3

[1] 7

> 2 * 5^2 - 10 * 5 # a comment

[1] 0

> 4 * sin(pi / 2)

[1] 4

> 0 / 0 # not defined (Not a Number)

[1] NaN

16 / 109


Arithmetic Operators, Functions, Values2

Operators, Functions, Values Description

∗, / multiplication, division+, - addition, subtraction%% modulo

max(), min() extremaabs() absolute valueround(), floor(), ceiling() round (up, down)sum(), prod() sum, productlog() logarithmsin(), cos() sinus, cosinus

pi the value of piInf, -Inf infinityNaN not defined (Not a Number)NA Not AvailableNULL empty set

2See R Development Core Team (2008): R: A Language and Environmentfor Statistical Computing. R Foundation For Statistical Computing, Vienna,Austria. URL http://www.R-project.org.

17 / 109


Assignments

How to ‘remember’ calculations

> x1 <- 3.25 # assign the value 3.25 to object x1

> x1

[1] 3.25

# please don’t use ‘=’ or ‘->’ for assignments

# <<- ‘special’ assignment - kept for later

Declaration and initialisation at once.

18 / 109


Assignments

How to ‘remember’ calculations

> x1 <- 3.25 # assign the value 3.25 to object x1

> x1

[1] 3.25

# please don’t use ‘=’ or ‘->’ for assignments

# <<- ‘special’ assignment - kept for later

Declaration and initialisation at once.

18 / 109


Print function - print()

What appears on screen? Try

> x1 <- 3.25

> (x1 <- 3.25)

> print(x1 <- 3.25)

> x1

19 / 109


Objects

R is an object oriented language

• Everything in R is an object (data, functions, etc.),

• objects have names (start with letter, upper-/lower-case),

• every object has a length(),

• every object is of a certain class - class(),

• for each class, there are special (generic) functions (e.g.print()),

• every object has a mode (data type) - mode(),

• an object may have attributes - attributes(),

• classes may inherit from other classes.

20 / 109


Find Help

Integrated help

• help.start()

• help(”cor”), ?cor

• help.search(”correlation”), ??correlation

21 / 109


Find Help

Internet

• check r-cran

• cran taskview

• cran discussions/mailing lists/examples

• manuals / vignettes

• R news/journal

• search for snippets - anything

22 / 109


Find Help

Books

• see literature first slide and many more.

23 / 109




24 / 109


Faithful - example

25 / 109


Workspace and working directory

Workspace - save what?

• workspace keeps your R-objects

• workspace is located in your RAM

• different environments in your workspace

26 / 109



Workspace - try out

> ls()

?

> search()

?

> ls(name="package:datasets")

?

27 / 109



Working directory - default folder for save/load etc. - try out

> getwd()

[1] "/Users/moritz"

> setwd("/Users/moritz/Documents/Teaching

/2019_2020/Urfist")

28 / 109


OverviewWhat is R?First stepsExamples

Data types and structuresData typesVectorMatrixListsData.framesConditional expressionsLoopsRead/WriteData cleaning/formattingData Merge/SelectionPlottingStatisticsBeyond this course

29 / 109


Data types - mode()

Description Example Data type

Basic data types (hierarchical)

empty set NULL NULLboolean TRUE logical

integers & reals 3.14 numericcomplex numbers 2.13+1i complex

characters and strings ”Hello” character

Composite data type

factors blue & red factor

30 / 109


Logical values in R

Try and explain

> 4 < 3

> (3+1) !=3

> -3<-2

> -3 < -2

> (3 >= 2) & (4 == (3+1))

> c(TRUE, FALSE) & c(TRUE, TRUE)

> x <- c(-4, -5, -1, 0, 2, 4, 5)

> x > 0

> sum(x > 0)

31 / 109


Logical values in R

Function, Operator, Value Description

==, != equal, not equal>, >= greater, greater equal<, <= smaller, smaller equal! not

&, && and (vector, non-vector)|, || or (vector, non-vector)xor() exclusive or

TRUE, FALSE true, false

32 / 109


Logical values in R

Try - any(), all(), which()

> x <- c(2, 0, 3, 1, 6, 9, 7)

> someTrue <- x > 4

> allTrue <- x >= 0

> any(someTrue)

> all(allTrue)

> any(allTrue)

> all(someTrue)

> !any(allTrue)

> all(!someTrue)

> all(!allTrue)

> which(someTrue)

> which(!someTrue) 33 / 109


Logical values in R

Not available - NA - conversion into other data types

> x <- NA

> mode(x)

>

> y <- c(3, x)

> mode(y)

34 / 109


Logical values in R

Not available - NA - is special

> # convert NA

> x <- NA

> mode(x)

> y <- c(3, x)

> mode(y)

> # comparison and calculation

> # - (if a part is unknown - result will be unknown)

> y > 2

> sum(y)

> # do not consider (remove) NA elements

> sum(y, na.rm=TRUE)

35 / 109


Numeric values in R

Try - mode(), typeof()

> (x <- pi)

[1] 3.141593

> mode(x)

[1] "numeric"

> typeof(x)

[1] "double"

> (y <- as.integer(x)) # information loss

[1] 3

> typeof(y)

[1] "integer"

36 / 109


Numeric values in R

Try - conversion

> is.character(y)

[1] FALSE

> x <- -1

> -1^0.5

[1] -1

> sqrt(as.complex(x))

[1] 0+1i

> (z <- as.character(y))

[1] "3"

> as.numeric("z")

[1] NA

Warning message:

NAs introduced by coercion

> (y <- as.numeric("-1"))

[1] -1

> y*5

[1] -5

37 / 109


Character values in R

There are just strings. Manipulating them is another story. Nextcourse.

"hello world"

a <- "hello world"

38 / 109


Factors in R

(ordered) factors

gulliver <- factor(c("dwarf","giant", "giant",

"giant","dwarf"))

gulliver <- factor(c("dwarf","giant", "giant",

"giant","dwarf"),

levels=c("giant","dwarf","gulliver"))

gulliver <- ordered(c("dwarf","giant", "giant",

"giant","dwarf"),

levels=c("dwarf","gulliver","giant"))

39 / 109


Factors in R

(ordered) factors

> class(gulliver)

[1] "ordered" "factor"

> str(gulliver)

Ord.factor w/ 3 levels "dwarf"<"gulliver"<..: 1 3 3 3 1

> mode(gulliver)

[1] "numeric"

> length(gulliver)

[1] 5

40 / 109




41 / 109


Data structures

Data structure is the representation of data, which defines how weoperate on the data.

Data structures in R

• vector

• matrix

• array

• list

• data.frame

For each data structure you have a function to declare & initialisean object, e.g. vector() .

42 / 109


Vector

characteristics of vectors

> x <- c(4.1, 4.3, 2, 54, 34) # c() for concatenate

>

> class(x)

[1] "numeric"

> mode(x)

[1] "numeric"

> typeof(x)

[1] "double"

> str(x)

num [1:5] 4.1 4.3 2 54 34

> is.vector(x)

[1] TRUE

> length(x)

[1] 5

> attributes(x)

NULL43 / 109


Vector


> # Only of one data type

> c(x, FALSE, TRUE)

[1] 4.1 4.3 2.0 54.0 34.0 0.0 1.0

> c(x, FALSE, TRUE, "Hello")

[1] "4.1" "4.3" "2" "54" "34" "FALSE" "TRUE" "Hello"

44 / 109


Vector


> x <- c(height=10, width=4)

> x

height width

10 4

> attributes(x)

$names

[1] "height" "width"

> names(x)

[1] "height" "width"

45 / 109


Vector

sequences

> 2:5

[1] 2 3 4 5

> seq(2,5,by=2)

[1] 2 4

> rep(2:4,times=3)

[1] 2 3 4 2 3 4 2 3 4

46 / 109


Vector

Calculate with vectors

> x <- c(2, 4, 8)

> x*2

[1] 4 8 16

> x-3

[1] -1 1 5

>

> x * c(3,2,1)

[1] 6 8 8

> x * c(3,2)

[1] 6 8 24

Warning message:

In x * c(3, 2) :

longer object length is not a multiple of shorter object length

47 / 109


Vector

Calculate with vectors

> t(2:4) %*% 1:3

[,1]

[1,] 20

> 2:4 %*% 1:3

[,1]

[1,] 20

> 2:4 %*% t(1:3)

[,1] [,2] [,3]

[1,] 2 4 6

[2,] 3 6 9

[3,] 4 8 12

48 / 109


Vector

Indexing vectors

> x

[1] 2 4 8

> x[2] # 2nd element by position

[1] 4

> x[c(FALSE,TRUE,FALSE)] # 2nd element with logical value

[1] 4

> names(x) <- c("Height","Length","Width")

> x["Length"] # access by name

Length

4

> x[-2] # without 2nd element

[1] 2 8

49 / 109


Vector

Indexing vectors

# vector x

x <- c(2, 4, 8)

# extract by boolean vector

idx <- x < 5

print(idx)

x[idx]

# overwrite

x[idx] <- 2

# extract by position vector

pos <- which(x < 5)

print(pos)

x[pos]

# overwrite

x[pos] <- 150 / 109


Vector

Indexing vectors

> x[3,2,1] # reverse order?

Error in x[3, 2, 1] : incorrect number of dimensions

> x[c(3,2,1)] # reverse order!

[1] 8 4 2

> x[] <- -2 # replace every element by -2

> x

[1] -2 -2 -2

> x <- -2 # overwrite x by -2

> x

[1] -2

51 / 109




52 / 109


Matrix

initialise a Matrix - try and explain

x <- 1:3

y <- 4:6

cbind(x,y)

rbind(x,y)

matrix(1:6, nrow=3, ncol=2)

matrix(1:6, nrow=3, byrow=TRUE)

matrix(1:6, ncol=2, byrow=TRUE)

matrix(1:6, nrow=3, ncol=4, byrow=TRUE)

53 / 109


Matrix

Operate on matrices - try and explain

A <- matrix(c(1,1,2,2), nrow=2, byrow=TRUE)

B <- t(A) # transpose

solve(A) # inverse

# elementwise operations

A * B

A / B

A - B

A^B

A == B

A == t(B)

any(A==B)

# matrix multiplication

A %*% B54 / 109


Matrix

Access matrix elements - exercise

>X <- matrix(1:6, nrow=3)

>rownames(X) <- c("length", "width", "height")

>colnames(X) <- c("shelf","chair")

• read length of shelf by [row pos., column pos.]

• read length of shelf by [row name, column name]

• read length of shelf by [row name, column pos.]

• overwrite length of shelf with 14

• get all dimensions of shelf

• get length of all mobiliar

• overwrite length of mobiliar by prior length times 10055 / 109


Matrix

Matrix features

> class(X)

[1] "matrix"

> typeof(X)

[1] "integer"

> mode(X)

[1] "numeric"

56 / 109


Matrix

Matrix features

> attributes(X)

$dim

[1] 3 2

$dimnames

$dimnames[[1]]

[1] "length" "width" "height"

$dimnames[[2]]

[1] "table" "chair"

57 / 109


Matrix

Matrix features

> str(X)

int [1:3, 1:2] 1 2 3 4 5 6

- attr(*, "dimnames")=List of 2

..$ : chr [1:3] "length" "width" "height"

..$ : chr [1:2] "table" "chair"

Matrices are special cases of vectors in R (the opposite in math ofcourse).

58 / 109


Array

Array - like matrix but possibly more dimensions

Avoid arrays.Don’t use arrays with more than 3 dimensions if not absolutelynecessary.Use map of array.

59 / 109




60 / 109


List

List - features

• very flexible,

• may contain different data types,

• also lists (recursive)

61 / 109


List

List - features

> (L1 <- list("this is a", 100, matrix(1:4,nrow=2),

list(TRUE,FALSE)))

[[1]]

[1] "this is a"

[[2]]

[1] 100

[[3]]

[,1] [,2]

[1,] 1 3

[2,] 2 4

[[4]]

[[4]][[1]]

[1] TRUE

[[4]][[2]]

[1] FALSE

62 / 109


List

List - features

> (L1 <- list(a="this is a", b=100, c=matrix(1:4,nrow=2),

truth=list(TRUE,FALSE)))

$a

[1] "this is a"

$b

[1] 100

$c

[,1] [,2]

[1,] 1 3

[2,] 2 4

$truth

$truth[[1]]

[1] TRUE

$truth[[2]]

[1] FALSE

63 / 109


List

List - features

> L1[[1]] # access by position

[1] "this is a"

> L1[["a"]] # access by name

[1] "this is a"

>

> L1$b # access by name (but different)

[1] 100

64 / 109


List

List - features

> L1[[1]] <- NULL

> L1

$b

[1] 100

$c

[,1] [,2]

[1,] 1 3

[2,] 2 4

$truth

$truth[[1]]

[1] TRUE

$truth[[2]]

[1] FALSE65 / 109




66 / 109


data.frames

data.frames

• the most typical structure for data sets

• data.frames are lists - but entries are vectors with same length

• data.frame()

67 / 109


Data.frame

initialise Data.frame

> shopping <- data.frame(Product=c("cheese","wine","bread"),

Unit=c("grams","bottles","loaf"), Amount=c(300,2,2))

>

> str(shopping)

’data.frame’: 3 obs. of 3 variables:

$ Product: Factor w/ 3 levels "bread","cheese",..: 2 3 1

$ Unit : Factor w/ 3 levels "bottles","grams",..: 2 1 3

$ Amount : num 300 2 2

68 / 109


Data.frame

Data.frame - access elements

> shopping[2,]

Product Unit Amount

2 wine bottles 2

> shopping[2,3]

[1] 2

> shopping[2,"Amount"]

[1] 2

> shopping$Amount[2]

[1] 2

69 / 109


Data.frame

Plot two regimes of old faithful

40 50 60 70 80 90 100

12

34

56

waiting

eruptions

70 / 109


Data.frame

Plot two regimes of old faithful (by subsetting data.frame)

# plot geysir data

plot(NA, xlim=c(40,100),ylim=c(1,6), xlab="waiting",

ylab="eruptions")

points(regime1[,"waiting"],regime1[,"eruptions"],col="red")

points(regime2[,"waiting"],regime2[,"eruptions"],col="blue")

points(rest[,"waiting"],rest[,"eruptions"],col="grey")

lines(x=c(70,70),y=c(0,7),lty=3)

lines(x=c(40,100),y=c(3.5,3.5),lty=3)

71 / 109


Data.frame

Plot two regimes of old faithful (by adding variables todata.frame)

# plot geysir data

plot(faithful$waiting,faithful$eruptions, col=faithful$color,

xlim=c(40,100),ylim=c(1,6),

xlab="waiting", ylab="eruptions")

lines(x=c(70,70),y=c(0,7),lty=3)

lines(x=c(40,100),y=c(3.5,3.5),lty=3)

72 / 109


OverviewWhat is R?First stepsExamplesData typesVectorMatrixListsData.frames

Control structuresConditional expressionsLoopsRead/WriteData cleaning/formattingData Merge/SelectionPlottingStatisticsBeyond this course

73 / 109


Conditional expressions

If...else

> x <- 1

> if(x > 0){

+ print("x is pos.")

+ }else{

+ print("x is neg.")

+ }

[1] "x is pos."

74 / 109



Ifelse

# vector-oriented, use for simple expr.

> x <- c(1,2,3,4)

> ifelse(x > 2, "A", "B")

[1] "B" "B" "A" "A"

75 / 109



switch

> switch(2, a=11, b=12, cc=13, d=14)

[1] 12

> (switch("c", a=11, b=12, cc=13, d=14))

NULL

> switch("cc", a=11, b=12, cc=13, d=14)

[1] 13

76 / 109


OverviewWhat is R?First stepsExamplesData typesVectorMatrixListsData.frames

Control structuresConditional expressionsLoopsRead/WriteData cleaning/formattingData Merge/SelectionPlottingStatisticsBeyond this course

77 / 109


Loops

for()

> for(i in 1:3){

+ print(i)

+ }

[1] 1

[1] 2

[1] 3

78 / 109


Loops

repeat

> i <- 0

> repeat{

+ i <- i+1

+ if(i < 3){

+ next # start next turn

+ }

+ print(i)

+

+ if(i == 3){

+ break # exit loop

+ }

+

+ }

[1] 379 / 109


Loops

while

> i <- 0

> while(i < 3){

+ print(i)

+ i <- i + 1

+ }

[1] 0

[1] 1

[1] 2

80 / 109


Loops in R

Short note on loops and efficiency

• R is executed by an interpreter (not compiled). Interpreterneeds to translate each loop into machine code.

• Vectorized code is typically faster (once translated, fetch,calculate write of each value ‘compressed’).

• maybe slower (if the size of the vector becomes too large),

• maybe difficult to read (if several objects are handled),

• maybe not possible (if sequential combination).

81 / 109


Excercise

• Exercise ‘R shaked’

• Exercise ‘Normal Distribution uncommented’

82 / 109


OverviewWhat is R?First stepsExamplesData typesVectorMatrixListsData.framesConditional expressionsLoops

Data HandlingRead/WriteData cleaning/formattingData Merge/SelectionPlottingStatisticsBeyond this course

83 / 109


Read/Write

Import/export, read/write

• R objects (load(), save())

• text-files (read.table() / write.table(), scan() / cat())

• operate on excel/access/data bases (e.g. RODBC, orRMySQL)

• SAS/SPSS/Stata (e.g. foreign)

See “R data import / export”.

84 / 109




85 / 109


Character string handling

Function Description

scan() / cat() Read from / print to file (or console)parse() / deparse() convert char into expression (and vice-versa)grep() regular expressionsnchar() number of chars in a stringstrsplit() / paste() (dis-)connect strings

See e.g. Sanchez, G. (2013) Handling and Processing Strings in RTrowchez Editions. Berkeley, 2013.http://www.gastonsanchez.com/Handling and Processing Stringsin R.pdf

86 / 109




87 / 109


Data Merge/Selection

How to join and split data.frames?

• merge(),

• rbind()

• cbind()

• subset()

• idx < − x=y; df[idx,c(”A”,”B”)]

Note: Merge/split of very large tables (millions of entries) better ina data base.

88 / 109


Data Merge/Selection

How to join and split data.frames?

• merge(),

• rbind()

• cbind()

• subset()

• idx < − x=y; df[idx,c(”A”,”B”)]

Note: Merge/split of very large tables (millions of entries) better ina data base.

88 / 109


OverviewWhat is R?First stepsExamplesData typesVectorMatrixListsData.framesConditional expressionsLoopsRead/WriteData cleaning/formattingData Merge/Selection


89 / 109


R plotting

How to make good graphs

1. Excel plots

2. R plots - various devices - pdf, ps etc.

3. TikZ plots (vector graphics in LaTeX) - TikzDevice - NEW

4. Plotly - interactive web-based graphs via the open sourceJavaScript graphing library NEWER

90 / 109


R plotting

Devices

• Device defines where to plot, similar to paper, pergament etc.

• the same graph gets different ‘look and feels’ on differentdevices

• (transparent) colours differ by device

91 / 109


R plotting - pdf-device

●

●

●

●●

● ● ●

●

●

●●

●●

●●●● ●●● ●

●● ●●●

2000 2003 2006 2009

1020

30Regional spread of Greek GDP

Year

k E

uro

per

capi

ta

92 / 109


Scientific plotting - tikzDevice

2000 2002 2004 2006 2008 2010

1015

2025

30

Regional spread of Greek GDP

Year

kE

uro

per

cap

ita

93 / 109


R plotting

Typical processStep 0: try out on the default device.

f <- paste(getwd(),"/GreekGDP.pdf",sep="") # define file

pdf(file=f, width=4, height=3) # open device

# plot into device

boxplot(gdp2[idx,"EUR_HAB"]/1000~gdp2[idx,"TIME"],notch=TRUE, xlab="Year", ylab="k Euro per capita",

main="Regional spread of Greek GDP")

dev.off() # close device

94 / 109


R plotting

Plot functions

• plot() - context dependent / generic function - depend on theclass to plot.

• add - points(), lines(), text() in the plot.

• barplot(), boxplot(), contour(), hist(), etc.

• plot functions from special packages

95 / 109


R plotting

• options in the function, e.g. cex

• options through par()

96 / 109


R plotting

Arguments frequently used in plot functions and par()Argument Description

axes (not) draw axes - later fine tunedcex size (multiply) of nodes and letterslty, lwd line type, line widthpch point symbolxlab, ylab axes labelsxlim, ylim axes dimensions

97 / 109


R plotting - regions

par()-ameters oma/mar3

3http://rgraphics.limnology.wisc.edu/rmargins sf.php98 / 109


R plotting - regions

par()-ameters mfcol/mfrow4

4http://rgraphics.limnology.wisc.edu/rmargins mfcol.php99 / 109


R plotting

General remarks

• Everything can be done,

• But may cost time to figure out how, and to code.

100 / 109




101 / 109


Descriptives

Checkoutsummary(), table(), prop.table(), cor(), rcorr(), etc.

102 / 109


Descriptives - free work

What’s the chance of a happy end? Use the ‘Titanic’ data inthe library ‘datasets’.

103 / 109


Descriptives - free work

Bring your project to the class room

YOUR PROJECT HERE.

104 / 109




105 / 109


Beyond this course

106 / 109


Inductive stats

Checkoutlm(), glm(), task view

107 / 109


Non-parametrics

Checkoutdensity() in R base, np-package, task view

108 / 109


Bayesian

CheckoutrStan package, task view

109 / 109