散点图基础2

散点图基础2

1. 拟合lm、loess、分组

library(ggplot2)
library(gcookbook)
sp<-ggplot(heightweight,aes(x=ageYear,y=heightIn))

#线性拟合lm
sp+geom_point()+stat_smooth(method=lm)  #默认95置信区间
sp+geom_point()+stat_smooth(method=lm,level = 0.99)#level设置置信区间
sp+geom_point()+stat_smooth(method=lm,se=FALSE)#se=false，不会添加置信区间
sp+geom_point()+stat_smooth(method=lm,se=FALSE,colour="red",linetype="dashed",size=2)#colour修改拟合线的颜色,线形（linetype）,粗细

#多项式拟合loess
sp+geom_point()+stat_smooth(method=loess)

#logis回归 ：不懂 见p77

#分组
sp<-ggplot(heightweight,aes(x=ageYear,y=heightIn,colour=sex))+
  geom_point()+
  scale_color_brewer(palette = "Set1")
sp+stat_smooth()

sp+stat_smooth(method = lm,fullrange = TRUE)#fullrange可以将拟合线绘制到右边界。loess不可以

2. 创建新的拟合函数

#lm建一个函数,predict（）计算预测值
library(gcookbook)
model<-lm(heightIn~ageYear+I(ageYear^2),heightweight)
model

xmin<-min(heightweight$ageYear)
xmax<-max(heightweight$ageYear)
a<-data.frame(ageYear=seq(xmin,xmax,length.out = 20))
a
a$heightIn<-predict(model,a)
a

sp<-ggplot(heightweight,aes(x=ageYear,y=heightIn))+geom_point(colour="grey40")
sp+geom_line(data=a,size=1)

summary(model)#查看参数值
#annotate（）手动添加标签
sp+annotate("text",label="r^2=0.42",x=16,y=55)

##3. 添加标签

#添加标签:手动annotate()，自动geom_text()
library(gcookbook)
cdata<-subset(countries,Year==2009&healthexp>2000)
cdata
sp<-ggplot(cdata,aes(x=healthexp,y=infmortality))+geom_point()
sp+annotate("text",label="Canada",x=4350,y=5.4)+annotate("text",label="USA",x=7400,y=6.8)

#把因子型或字符串型赋值给label，默认将标签中心置于xy坐标位置，默认size=5
sp+geom_text(aes(label=Name),size=4)

#对标签上下左右调整,vjust=0，文本基线会与数据点对齐，vjust=1，文本顶部与数据点对齐。hjust=0，左对齐，hjust=1，右对齐。或者对x，y增加或者减小。

sp+geom_text(aes(label=Name),size=4,hjust=0,vjust=1)
sp+geom_text(aes(x=healthexp+100,y=infmortality+0.2,label=Name),size=4)

#给几个点自动添加标签：增加一列只包含这几个点的新的标签列 .%in%判断前面一个向量内的元素是否在后面一个向量中，返回布尔值。
cdata$name1<-cdata$Name
abx<-cdata$name1 %in% c("Canada","Ireland","Japan")
abx
cdata$name1[!abx]<-NA
cdata
sp<-ggplot(cdata,aes(x=healthexp,y=infmortality))+geom_point()+
  geom_text(aes(x=healthexp+100,y=infmortality+0.2,label=name1),size=4)+xlim(2000,10000)
sp

##4.气泡图，scale_size_area

#绘制气泡图
#size=GDP,GDP赋值给半径
p<-ggplot(cdata,aes(x=healthexp,y=infmortality,size=GDP))+geom_point(shape=21,colour="black",fill="cornsilk")
p
#GDP赋值给点的面积
p+scale_size_area(max_size = 15)