数据可视化

---

一、散点图

代码如下（示例）：

#游戏数据可视化
# 绘制散点图
w <- read.csv("可视化数据.csv", header = F, 
              as.is=T)
str(w)
head(w)
data.names <- c("日期","活跃用户","新增用户","付费率",
                "付费用户","首次付费用户","ARPPU","ARPU")
colnames(w) <- data.names
str(w)
head(w)
#单指标数据可视化
summary(w)
w$new.date <- paste(substr(w$日期,6,7),substr(w$日期,9,10),sep = "/")
head(w)
barplot(height = w$新增用户, names.arg = w$new.date)
#还可以进行哪些单指标分析？通过分析单指标，得出相对应的结论。

二、散点图

#双指标数据可视化
plot(w$活跃用户~w$新增用户,col="violetred2",pch=16,
     main="活跃用户 vs 新增用户散点图")
#添加线性拟合直线
abline(lm(w$活跃用户~w$新增用户),col="blue",lwd=2,lty=2)
#还可以进行哪些双指标分析？通过分析双指标，得出相对应的结论。

#三指标数据可视化
plot(w$活跃用户~w$新增用户,col="violetred2",pch=16,
     cex=w$付费率,
     main="活跃用户 vs 新增用户气泡图")

三、趋势分析

#同比、环比
data <- data.frame(pre=c(113,134,123,145,137,196,187),
                   now=c(129,122,134,149,146,215,208));
ylim.max <- 550
col = c("azure4","brown4")
##将主副标题放到barplot函数里，进行设置
##font=3,表示字体为“斜体”
##cex=1.5,表示字体放大1.5倍
barplot(as.matrix(rbind( data$pre,data$now)),
        beside=TRUE,ylim=c(0,ylim.max),col=col,axes=F,
        main=list( "本周pv趋势分析图",cex=1.5,col="red",font=3),
        sub=paste("范围：2020.09.28--2020.10.04","\n","网站板块：军事科技"),
        ylab="网站日页面浏览量pv"
)
axis(2)

##画图例
text.legend = c("上周pv","本周pv","pv同比增长","pv环比增长")
col2 <- c("black","blue")
legend("topleft",pch=c(15,15,16,16), legend=text.legend,cex=0.8,col=c(col,col2),bty="n",
       horiz=F)

##画刻度标记
text.x <- c("周一","周二","周三","周四","周五","周六","周日")
axis(side=1, c(2,5,8,11,14,17,20), labels=text.x,tick=TRUE, cex.axis=0.75)
axis(4,at=seq(from=250,length.out=7,by=40),labels=c("-60%","-40%","-20%","0","20%","40%","60%"))

##添加“同比增长曲线”和“环比增长曲线”
##同比增长 = (now[t]-pre[t])/pre[t]
same.pre.growth <- (data$now-data$pre)/data$pre
##环比增长 = (now[t]-now[t-1])/now[t-1]
ring.growth <- c(NA,diff(data$now)/data$now[1:(length(data$now)-1)])
a<-200;  b<-370;
lines(c(2,5,8,11,14,17,20), a*same.pre.growth+b,type="b",lwd=2)
lines(c(2,5,8,11,14,17,20), a*ring.growth+b,type="b",lwd=2,col="blue")

##在同比和环比曲线上添加文字
j <- 1
for(i in 1:length(data[,1])){ #在bar上画数值
  text(3*i-1, a*same.pre.growth[i]+b-5, paste(round(same.pre.growth[i]*10000)/100,
                                              "%",sep=""));
  j <- j+1;
  text(3*i-1, a*ring.growth[i]+b+5, paste(round(ring.growth[i]*10000)/100,
                                          "%",sep=""),col="blue");
  j <- j+2;
}
##在pv柱状图上添加文字
j <- 1
for(i in 1:length(data[,1])){
  text(j+0.5,data$pre[i]+10, data$pre[i], col="deepskyblue4");
  j < j+1  #组内间距为1
  text(j+1.5,data$now[i]+10, data$now[i], col="deepskyblue4");
  j <- j+3 #组外间距为3
}

四、相关性分析

logindata <- read.csv("logindatatest.csv", 
                      header = T,  fileEncoding = "UTF-8")
logindata <- read.csv("abalone.csv", 
                      header = T,  fileEncoding = "UTF-8")
str(logindata)
nrow(logindata)
dim(logindata)
library(caret)
dmy<-dummyVars(~.,data=logindata)
dmyTsrf<-data.frame(predict(dmy,newdata=logindata))
dim(dmyTsrf)
str(dmyTsrf)
log.cor <- cor(dmyTsrf, use="pairwise.complete.obs")
#install.packages("corrplot")
library(corrplot)
corrplot.mixed(log.cor[1:9,1:9], lower = "ellipse", 
               upper = "number",
               tl.pos="lt",diag="u")

# 漏斗分析
# 导入新手教程关键路径用户数 
funnel <- read.csv("新手教程路径留存人数统计.csv",
                   header = T,  fileEncoding = "UTF-8")
funnel <- read.csv("glass.csv",
                   header = T,  fileEncoding = "UTF-8")
# 查看funnel数据集
funnel
# 绘制漏斗图
# 对数据进行降序排序
funnel_order <- funnel[order(funnel$用户数),]
# 绘制漏斗图
barplot(funnel_order$用户数,horiz = T,axes=F,border=F,space=0.5,
        col="steelblue1",xlim=c(-700,700),
        main="新手教程关键路径漏斗图")
barplot(-funnel_order$用户数,horiz=T,add=T,border=F,space=0.5,
        axes=F,col="steelblue1")
# 增加步骤名称
text(x=rep(-600,6),y=seq(1,8.5,length.out = 6),
     labels=funnel_order$事件行为,cex=0.8,font=3,col="black")
# 增加总体的转化率
text(x=rep(0,6),y=seq(1,8.5,length.out = 6),
     labels=paste0(round(funnel_order$用户数*100/max(funnel_order$用户数),2),"%"),
     cex=0.8,font=3,col="violetred3")
# 增加上一步的转化率
text(x=rep(max(funnel$用户数),6),y=seq(2,7.5,length.out = 5),
     labels=paste0(round(funnel_order$用户数[1:5]/funnel_order$用户数[2:6]*100,2),"%"),
     cex=0.8,font=4,col="red3")