R for beginners

1

R for beginners

Emmanue

l Paradis

1 W^ h



at^ i



s R ? 3

2 The few t^ hings t^ o know b^ efore st^art^ ing 5

2



.1 T h e o p e^ra^ t^o r^ <- 5^

2.2 Lis^ting^ a^ nd^ d^ e le t^ing^ t^he o b^{ }je^ct^s^ in me mo ry^ 5^

2.3^ The o n-line he lp 6^

3



Dat^ a wit^h R 8

3



.1 The ‘o b



j



e

 c^ t^s^ ’ 8

3

 .2^ R e a^ d^ in g^ d^ a^ t^a^ fr^o m^! fil^"e s^ 8

3



.3^ S^# a^{ $}vin g^ d^ a^ t^a^ 10^%

3

 .4^& G^' e n e r^a^ t^in g^ d^ a^ t^a^ 11

3



.4.1 Re g^ u⁽ la^ r s^ e q⁾ u⁽ e nc^ e s^ 11 3



.4.2 Ra^ nd^ o m s^ e q^{) (}ue nc^ e s^ 13^

3



.5^ Ma^ nip u⁽ la^ t^ing^ o b^{ }je c^ t^s 13^ 3



.5^ .1 A^* c^ c^ e s^ s^in g^ a^ p a^ r^t^ic^ u⁽ l^"a^ r^ v^$ a^ l^"u⁽ e o f a^ n o b^{ }je c^ t^ 13^

3



.5^ .2^ A^* r^it^h m^! e t^ic^ s^ a^ n d^ s^ i^!mp l^"e fu⁽ n^ct^ion s^ 14^&

3



.5^ .3^ M⁺ a^ t^rix^, c^ om^! p u⁽ t^at^io n 16^

4

-

Gr^. ap^{/ }h^ics wi^th^ R 18

4

&

.1 M⁺ a^ n a^{ }gin^g g^ r^a^ p h ic^ w⁰ in d^ o w⁰ s^ 18 4.1.1 O¹ p e ning^ s^ e v^$ e ra^ l g^ ra^ p hic^ w⁰ ind^ o w⁰ s^ 18 4.1.2 Pa^ rt^it^io ning^ a^ g^ ra^ p hic^ w⁰ ind^ o w⁰ 18

4.2 G^' ra^ p hic^ fu⁽ nc^ t^io ns^ 19²

4

&

.3^ L³ ow⁰ -l^"ev^$ e l^" p l^"o t^tin g^ c^ o m^! m^! a^ n d^ s^ 2^ 0^%

4

&

.4^& G^' r^a^ p h ic^ p a^ r^a^{ !}me^te r^s^ 2^ 1

5 St^ at^ist^ical⁴ an⁵ al⁴y⁶ ses wi^t^h^ R 23

6 The p^/ rogramming langu⁷ age R 26

6



.1 Lo o p s^ a^ nd



c^ o nd



it^io na^ l e xe c^ u⁽ t^io ns^ 26



6



.2^ W⁸ r^i

t

i

n

 g^ y^ o u⁽ o w⁰ n fu⁽ n c^ t^i

o

 n s^ 2^ 7⁹

7

:

H^; ow t^o go f

<

ar^{. }th



er^. wi



t

h



R ? 30

8 Index 31

2

T

=

h



e goal⁴ of^{< }t^he p^/ r^. esen⁵ t^ docu⁷ m^> en^{5 }t^is t^ o gi^?ve a st^ ar^{. }t^i⁵ng p^/ oi^5n^tf^< or^. p^/ eop^/ l⁴e n⁵ ewl⁴⁶y^i⁵n^ter^. est^ed i^n⁵ R. I^@ t

 ried t^o simp^/ lify⁶ as mu⁷ ch as I cou⁷ ld t^ he exp^/ lanat^ ions t^o make t^ hem u⁷ nderst^ andab^ les b^ y⁶ all, while giv^? ing u⁷ sefu⁷ l det^ails, somet^ imes wit^ h t^ab^ les. Commands, inst^ ru⁷ ct^ ions and examp^/ les are writ^ten in ^{CouA rier} font^.

I t^ hank J^{B 7}ulien Clau⁷ de, Christ^ op^/ he Declercq, Friedrich Leisch and Mat^ hieu⁷ Ros for t^ heir comment^ s and su⁷ ggest^ ions on an earlier v^? ersion of t^his docu⁷ ment^. I am also grat^ efu⁷ l t^ o all t^he memb^ ers of t^ he R Dev^? elop^/ ment^ Core Team for t^ heir considerab^ le effort^ s in dev^? elop^/ ingRand animat^ ing t^ he discu⁷ ssion list^ ‘r-help^/ ’. Thanks also t^ o t^ he R u⁷ sers whose qu⁷ est^ ions or comment^ s help^/ ed me t^o writ^ e “R for b^ eginners”.

© 2000, E^{C >}m^>man^{5 7}uel

4

P^D ar^. adi



s (20 oct^ob



r

. e 2000)

3

1 W

ha

t

is

R ?

R is a st^at^ist^ical analy⁶ sis sy⁶ st^ em creat^ed b^ y⁶ Ross Ihaka & Rob^ ert^ Gent^ leman (1996, J^I . Com^{J K}put. Gra^{L K}ph^M . S^N ta^L t., 5: 299-314). Ris b^ ot^ h a langu⁷ age and a soft^ware; it^ s most^ remarkab^ le feat^7ures are:

• an⁵ ef

<

f

<

ect^ i



v

? e dat^a h



an⁵ dl

4

i



n

5 g an⁵ d st^ or^. age f

<

aci



l

4

i



t

 y⁶ ,

• a su⁷ i^t^e of^< op^/ er^. at^ or^. s f^< or^. cal⁴ cu^{7 4}lat^ion⁵ s on⁵ ar^{. .}ray⁶ s, m^> at^.r^ices, an⁵ d ot^ h^ er^. com^{> /}p⁴lex^O op^/ er^. at^ion⁵ s,

• a large, coherent^ , int^ egrat^ed collect^ ion of t^ ools for st^ at^ ist^ ical analy⁶ sis,

• nu⁷ merou⁷ s grap^/ hical facilit^ies which are p^/ art^icu⁷ larly⁶ flexib^ le, and

• a simp^/ le and effect^ iv^? e p^/ rogramming langu⁷ age which inclu⁷ des many⁶ facilit^ ies.

R ^is a l⁴ an⁵ gu⁷ age con⁵ si^ der^. ed as a di^ al⁴ ect^ of^{< }t^he l⁴an⁵ gu⁷ age S cr^. eat^ ed b^ y⁶ t^he A^{P =}T&T^{= Q}Bel⁴⁴l L

R ab^ or^. at^or^. i^es. S i^ s av^? ai^ l⁴ ab^ l⁴ e as t^ h^ e sof^< t^war^. e S-P^D L^R U^S S com^> m^> er^. ci^ al⁴ i^ zed b^ y⁶ M^T at^ h^ Sof^< t^ (see h



t

 t^ p^/ :/^U /^U www.sp^/ l⁴ u⁷ s.m^> at^ h^ sof^< t^ .com^> /^U f^< or^. m^> or^. e i^n⁵ f^< or^. m^> at^ i^ on⁵ ). T⁼ h^ er^. e ar^. e i^ m^> p^/ or^. t^an⁵ t^s di^ f^< f^< er^. en⁵ ces i^ n⁵ t

 he concep^/ t^ ions of R and S, b



u

7 t^ t^ hey⁶ are not^ of int^ erest^ t^ o u⁷ s here: t^ hose who want^ t^ o know more on t^ his p^/ oint^ can read t^he p^/ ap^/ er b^{ 6}y Gent^ leman & Ihaka (1996) or t^ he R-FAQ (ht^t^/p:/^UU/cran.r-p^/ roj^V ect^ .org/^Udoc/^UFAQ/^UR-FAQ.ht^ ml), a cop^/ y⁶ of which is alse dist^ rib^{ 7}u^ted wit^ h t^ he soft^ ware.

R ^is f^{< .}reel⁴y⁶ di^ st^.r^i^b⁷u^ted on⁵ t^h^ e t^ er^{. >}ms of^{< }t^he GN^W U^{S D}P⁷u^b⁴l^ic L^{R }icen⁵ ce of^{< }t^he F^{X .}ree Sof^<twar^. e F

X

ou^{7 5}ndat^ion⁵ (f^< or^{. >}mor^. e i^5n^<for^{. >}mat^ion⁵ : h^{ }t^t^/p:/^U/^Uwww.gn^{5 7}u.or^. g/^U ); i^ts dev^? el⁴ op^{/ >}men^{5 }tan⁵ d di^ st^.r^i^b⁷u^t^ion⁵ ar^. e carried on b



y

6 sev^? eral st^ at^ ist^ icians known as t^ he R Dev^? elop^/ ment^ Core Team. A key⁶ -element^ in t

 his dev^? elop^/ ment^ is t^ he Comp^/ rehensiv^? e R Archiv^? e N^W et^ work (CRAN^W ).

R is av^? ailab^ le in sev^? eral forms: t^ he sou⁷ rces writ^t^ en in C (and some rou^{7 }tines in Fort^ ran77) ready^{6 }to b^ e comp^/ iled, essent^ ially⁶ for U^S nix and Linu⁷ x machines, or some b^ inaries ready⁶ for u⁷ se (af^{< }ter^. a v^? er^. y⁶ easy⁶ i^n⁵ st^al⁴ l⁴at^ i^on⁵ ) accor^. di^5ng t^ o t^h^ e f^< ol⁴⁴lowi^5ng t^ ab^{ 4}le.

Arc^Y hit^Z e^[ c^Y t^Zu^\ re^[ O^] p^{^} e^[ ra^_ t^Zing^` s^a y^b s^a t^Ze^[ m(^cs^a )^d I

e

n

 t^e l^" W⁸ in d^ o w⁰ s^ 9² 5^ /^f9² 8 /^fN^g T 4^& .0^% /^f2^ 0^% 0^% 0^%

L

3

i

n

 u⁽ x^, (^hDⁱ e b^ ia^ n 2^ .2^ ,^j M⁺ a^ n d^ r^a^ k^k e 7⁹ .1,^j R e d^ H^l a^ t^

6



.x,^j S^# u⁽ S^# e 5^ .3^ /^f6.4/^f97.0^% )^m P

n

P

n

C

o

M

+

a

 c^ O¹ S^#

Linu⁽ xPPC^o 5^ .0^% Alp ha^ S^# y^ s^te ms^ Dig^ it^a^ l U^p nix 4.0^%

Linu⁽ x (^hRe d^ Ha^ t^ 6^ .x)^m S

#

p

a^ rc^ Linu⁽ x (^hRe d



Ha^ t^ 6



.x)^m

T

=

h



e f^{< }i⁴les t^ o i^ n⁵ st^al⁴ l⁴ t^hese b^ i^n⁵ ar^. i^es ar^. e at^ h^ t^t^p^/ :/^U/^U cr^. an⁵ .r^. -p^{/ .}roj^V ect^ .or^. g/^Ub^i⁵n^U/(ex^O cep^{/ }t^<for^{. }t^he M^T aci^5n^tosh^ v

? ersion ¹) where y⁶ ou⁷ can find t^ he inst^ allat^ ion inst^ru⁷ ct^ions for each op^/ erat^ing sy⁶ st^ em as well.

R is a langu⁷ age wit^ h many⁶ fu⁷ nct^ ions for st^ at^ ist^ ical analy⁶ ses and grap^/ hics; t^ he lat^t^er are v

? isu⁷ alized immediat^ ely⁶ in t^heir own window and can b^ e sav^? ed in v^? ariou⁷ s format^ s (for ex^O am^{> /}p⁴le, j^{V /}pg, p^{/ 5}ng, b^ m^> p^/ , ep^/ s, or^. wm^> f^< u⁷ n⁵ der^{. }W^in⁵ dows, p^/ s, b^{ >}mp^/ , p^{/ }ict^ ex^{O 7}u⁵nder^{. S}U⁵n^i^Ox). T⁼ h^ e r

. esu⁷ l⁴ t^ s f^< r^. om^> a st^at^i^st^ i^ cal⁴ an⁵ al⁴ y⁶ si^ s can⁵ b^ e di^sp^/ l⁴ay⁶ ed on⁵ t^ h^ e scr^. een⁵ , som^> e i^ n⁵ t^ er^. m^> edi^ at^ e r^. esu⁷ l⁴ t^ s (P^q -v^? al⁴⁷ues, r^. egr^. essi^ on⁵ coef^{< <}f^ici^ en^{5 }ts) can^{5 }be wr^{. }i^t^ten^{5 }i⁵na f^{< }i⁴le or^{. 7}used i^5nsu^{7 }bsequ⁷ en^{5 }tan⁵ al⁴⁶yses. T^{= }heR l

4

an⁵ gu⁷ age al⁴l⁴ows t^h^ e u⁷ ser^. , f^< or^. i^n⁵ st^ an⁵ ce, t^ o p^{/ .}rogr^. am^> l⁴oop^/ s of^< com^{> >}man⁵ ds t^ o su⁷ ccessi^?vel⁴⁶yan⁵ al⁴⁶yse

1 T h e M⁺ a^ c^ in t^o s^ h p o r^t^ o f R^r h a^ s^ j^ u⁽ s^ t^ b^ e e n fin is^ h e d^ b^ y^ S^# t^efa^ n o I^eac^ u⁽ s^ <^{s }j^ag^ o@m^! c^ l^"in k^k .it^>^t ,^j a^nd^ s^ h o u⁽ l^"d b^ e a^{ $}va^il^"a^ b^ l^"e s

 o o n o n C^o R A^* N^g .

4

sev^? eral dat^ a set^ s. It^ is also p^/ ossib



le t^ o comb



ine in a single p^/ rogram different^ st^ at^ ist^ ical f

<

u

7 n⁵ ct^ i^ on⁵ s t^o p^/ er^. f^< or^. m^> m^> or^. e com^> p^/ l⁴ ex^O an⁵ al⁴ y⁶ ses. T⁼ h^ e R u⁷ ser^. s m^> ay⁶ b^ en⁵ ef^< i^ t^ of^< a l⁴ar^. ge n⁵ u⁷ m^> b^ er^. of^<

r

. ou⁷ t^ i^ n⁵ es wr^. i^ t^ t^ en⁵ f^< or^. San⁵ d av^? ai^ l⁴ ab^ l⁴ e on⁵ i^n⁵ t^er^. n⁵ et^ (f^< or^. ex^O am^> p^/ l⁴e: h^ t^ t^p^/ :/^U/^Ust^at^.cm^> u⁷ .edu⁷ /^US/^U), m^> ost^ of^<

t

 h^ ese r^. ou⁷ t^i^n⁵ es can⁵ b^ e u⁷ sed di^ r^. ect^ l⁴ y⁶ wi^ t^ h^ R.

A

P t^ f^< i^r^. st^, R cou⁷ l⁴ d seem^> t^oo com^> p^/ l⁴ ex^O f^< or^. a n⁵ on⁵ -sp^/ eci^ al⁴ i^ st^ (f^< or^. i^n⁵ st^an⁵ ce, a b^ i^ ol⁴ ogi^ st^ ). T⁼ h^ i^ s m^> ay⁶ n

5 ot^ b^ e t^ r^. u⁷ e act^ u⁷ al⁴ l⁴ y⁶ . I^@ n⁵ f^< act^ , a p^/ r^. om^> i^ n⁵ en⁵ t^ f^< eat^ u⁷ r^. e of^< R i^ s i^ t^ s f^< l⁴ ex^O i^ b^ i^ l⁴ i^ t^ y⁶ . W^ h^ er^. eas a cl⁴ assi^ cal⁴ sof^{< }twar^. e (SA^P S, SP^D SS, St^at^i^st^i^ca, ...) di^ sp^{/ 4}lay⁶ s (al⁴m^> ost^ ) al⁴⁴l^t^he r^. esu^{7 4}lt^s of^< an⁵ an⁵ al⁴⁶ysi^ s,R st^or^. es t

 hese resu⁷ lt^ s in an obj^u ect, so t^ hat^ an analy⁶ sis can b^ e done wit^h no resu⁷ lt^ disp^/ lay⁶ ed. The u⁷ ser may⁶ b^ e su⁷ rp^/ rised b^{ 6}y^thu⁷ s, b^{ 7}u^tsu⁷ ch a feat^7ure is v^? ery^{6 7}usefu⁷ l. Indeed, t^ he u⁷ ser can ext^ ract^ only⁶ t

 he p^/ art^ of t^ he resu⁷ lt^ s which is of int^ erest^ . For examp^/ le, if one ru⁷ ns a series of 20 regressions an⁵ d wan^{5 }ts t^ o com^{> /}par^. e t^he di^<f^<fer^. en⁵ t^{ .}regr^. essi^on⁵ coef^< f^< i^ci^en⁵ t^s, R can⁵ di^ sp^{/ 4}lay⁶ on^{5 4}l⁶y^t^he est^i^>mat^ ed coef^{< <}f^ici^ en^{5 }ts: t^h⁷us t^he r^. esu⁷ l⁴t^ s wi^l⁴⁴l^tak^v e 20 l⁴ i^n⁵ es, wh^ er^. eas a cl⁴ assi^ cal⁴ sof^{< }twar^. e cou⁷ l⁴d wel⁴ l⁴ op^/ en⁵ 20 r^. esu^{7 4}l^ts wi^ n⁵ dows. On⁵ e cou^{7 4}ld ci^te m^> an^{5 6}y ot^h^ er^. ex^O am^> p^/ l⁴es i^l⁴⁴l⁷ust^.rat^i⁵ng t^he su^{7 /}per^. i^or^. i^t^y⁶ of^< a sy⁶ st^ em^> su⁷ ch



asR com^{> /}par^. ed t^ o cl

4

assi



cal

4

sof

<

t

 war^. es; I^@ h



op^/ e t^h



e r^. eader^. wi



l

4

l

4

b



e con^{5 ?}vi



n

5 ced of

<

t

 h



i



s af^{< }ter^. r^. eadi^5ng t^h^ i^s docu^{7 >}men^{5 }t.