語言擴展包數(shù)據(jù)加工廠plyr視頻教程3課

R七種之數(shù)據(jù)plyr講師R七種之數(shù)據(jù)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)plyr——第3周法律【】和幻燈片為煉數(shù)成金網(wǎng)絡(luò)課程的教學資料，所有資料只能在課程內(nèi)使用，不得在課程以外范圍散播，違者將可能被法律和經(jīng)濟責任。課程詳情

煉數(shù)成金培訓http:

DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師關(guān)注煉數(shù)成金企業(yè)提供全面的數(shù)據(jù)價值資訊，涵蓋商業(yè)智能與數(shù)據(jù)分析、大數(shù)據(jù)、企業(yè)信息化、數(shù)字化技術(shù)等，各種高性價比課程信息，趕

出您

機關(guān)注吧！DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例分析baseball案例ozone案例DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師Baseball數(shù)據(jù)集職業(yè)棒球球手15年或以上的擊球記錄Baseball數(shù)據(jù)集包含了全id：棒球選手year：記錄的年份rbi：runsbatted

in，該球員在某一賽季的擊球的跑壘得分ab：number

of

times

at

bat，擊球次數(shù)head(baseball,5)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball求某一選手職業(yè)生涯時間baberuth

<-

subset(baseball,

id

==

"ruthba01")baberuth

<-

transform(baberuth,

cyear

=

year

-

min(year)+

1)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball求所有選手的職業(yè)生涯時間baseball

<-

ddply(baseball,.(id),

transform,cyear

=

year

-

min(year)

+

1)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball生成所有人的rbi/ab的時間序列圖并保存到pdf中baseball

<-

subset(baseball,

ab

>=

25)xlim

<-

range(baseball$cyear,

na.rm=TRUE)ylim

<-

range(baseball$rbi

/

baseball$ab,

na.rm=TRUE)plotpattern

<-

function(df)

{qplot(cyear,

rbi

/

ab,

data

=

df,

geom

=

"line",

xlim

=xlim,

ylim

=

ylim)}pdf("paths.pdf",

width

=

8,

height

=

4)d_ply(baseball,

.(reorder(id,

rbi

/

ab)),

failwith(NA,

plotpattern),.print=TRUE)dev.off()DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseballDATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball對ruthba01做線性回歸model

<-

function(df)

{lm(rbi

/

ab

~

cyear,

data

=

df)}model(baberuth)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball對所有選手做線性回歸bmodels

<-

dlply(baseball,

.(id),

model)rsq

<-

function(x)

summary(x)$r.squaredbcoefs

<-

ldply(bmodels,

function(x)

c(coef(x),

rsquare

=

rsq(x)))names(bcoefs)[2:3]

<-

c("intercept",

"slope")head(bcoefs,5)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball查看線性方程擬合效果hist(bcoefs$rsquare,breaks=20,col="black")DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：baseball找出擬合得較好的模型baseballcoef

<-

join(baseball,

bcoefs,

by

=

"id",type="inner")subset(baseballcoef,

rsquare

>

0.999)$idmatch_df(baseball,subset(baseballcoef,

rsquare

>

0.999),on="id")DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone數(shù)據(jù)集介紹：Ozone數(shù)據(jù)集是一個三維數(shù)組，記錄了24×24個空間網(wǎng)格內(nèi)，從月，共72個時間點上，

洲每月的平均臭氧水平。前兩維分別表示緯度和經(jīng)度，第三維表示時間。年月到年DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone研究位置（1,1）在不同時點臭氧水平的變化。方法一：按照所有時間點畫出折線圖value

<-

ozone[1,

1,

]time

<-

1:72

/

12plot(c(1:72),value,main=NULL,xlab="time",ylab="value",type="l")box(bty="l")grid(nx=NA,ny=NULL,lty=1,lwd=1,col="gray")DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone方法二：對不同年份按時點畫出折線圖plot(value[1:12],type="b",pch=19,lwd=2,xaxt="n",col="black",xlab="month",ylab="value")axis(1,at=1:12,labels=c("Jan",

"Feb",

"Mar",

"Apr",

"May","Jun",

"Jul",

"Aug",

"Sep",

"Oct",

"Nov",

"Dec"))lines(value[13:24],col="red",type="b",pch=19,lwd=2)lines(value[25:36],col="orange",type="b",pch=19,lwd=2)lines(value[37:48],col="purple",type="b",pch=19,lwd=2)lines(value[49:60],col="blue",type="b",pch=19,lwd=2)lines(value[61:72],col="green",type="b",pch=19,lwd=2)legend("bottomright",legend=c("1995","1996","1997","1998","1999","2000"),lty=1,lwd=2,pch=rep(19,6),col=c("black","red","orange","purple","blue","green"),ncol=1,bty="n",cex=1.2,text.col=c("black","red","orange","purple","blue","green"),inset=0.01)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：OzoneDATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone對位置（1,1），以月份為自變量，做一個穩(wěn)健線性回歸month.abbr

<-

c("Jan",

"Feb",

"Mar",

"Apr",

"May","Jun",

"Jul",

"Aug",

"Sep",

"Oct",

"Nov",

"Dec")month

<-

factor(rep(month.abbr,

length

=

72),

levels

=

month.abbr)year

<-

rep(1:6,

each

=

12)library("MASS")deseas1

<-

rlm(value

~

month

-

1)summary(deseas1)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone畫出殘差圖DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone觀察穩(wěn)健回歸模型的系數(shù)coef(deseas1)穩(wěn)健回歸方程DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone對所有24×24=576個位置做穩(wěn)健線性回歸deseasf

<-

function(value)

rlm(value

~

month

-

1,

maxit

=

50)models

<-

alply(ozone,

1:2,

deseasf)用failed

無法實現(xiàn)穩(wěn)健線性回歸的位置failed

<-

laply(models,

function(x)

!x$converged)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone提取所有模型的系數(shù)和殘差coefs

<-

laply(models,

coef)dimnames(coefs)[[3]]

<-

month.abbrnames(dimnames(coefs))[3]

<-

"month"deseas

<-

laply(models,

resid)dimnames(deseas)[[3]]

<-1:72names(dimnames(deseas))[3]

<-"time"coefs是一個三維數(shù)組，記錄了所有24×24個位置中每個位置的12個系數(shù)>dim(coefs)[1]

24

24

12deseas是一個三維數(shù)組，記錄了所有24×24個位置中每個位置的72個殘差dim(deseas)[1]

24

24

72DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：OzoneDATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozonecoefs_df

<-

melt(coefs)head(coefs_df)coefs_df

<-

ddply(coefs_df,

.(lat,

long),

transform,

avg

=

mean(value),std

=

value/max(value))head(coefs_df)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozonedeseas_df

<-

mehead(deseas_df)eseas)DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozonecoef_limits

<-

range(coefs_df$value)coef_mid

<-

mean(coefs_df$value)monthsurface

<-

function(mon){df

<-

subset(coefs_df,

month

==mon)qplot(long,

lat,

data

=

df,

fill

=

value,

geom="tile")

+scale_fill_gradient(limits

=

coef_limits,low

=

"lightskyblue",

high

=

"yellow")}DATAGURU專業(yè)數(shù)據(jù)分析社區(qū)R七種之數(shù)據(jù)plyr講師案例：Ozone以下兩幅圖分別展現(xiàn)了1月與7月的區(qū)域差異。months