R可視化lend_club 全球最大的P2P平臺數據75W條

lend_club 全球最大的P2P平臺2007~2012年貸款數據百度雲下載。
此文章基於R語言作簡單分析。

rm(list=ls())  #清除變量
gc()           #釋放內存

• step1
  考慮到後續分析
  將數據導入sqlserver,用到SSIS
  如圖

**此處有坑

• step2
  鏈接sqlserver,並將數據讀入R。

library(RODBC)
con<-odbcConnect("LI")   # LI 是本地數據庫，con~connect 是本地鏈接

RODBC Connection 2
Details:
  case=nochange
  DSN=LI
  UID=
  Trusted_Connection=Yes
  APP=RStudio
  WSID=LIYI-PC

lend_club1<-sqlQuery(con,"SELECT sum([Amount Requested]) as sumamount
      ,[Application Date] as date_1
      ,[year]
               ,substring(convert(varchar(12),[Application Date],111),6,5) as month_day
               FROM [liyi_test].[dbo].[lend_club]
               group by [year],substring(convert(varchar(12),[Application Date],111),6,5),[Application Date]
               order by [year],[month_day]")

head(lend_club1)
sumamount     date_1 year month_day
1      2000 2007-05-26 2007     05/26
2     47400 2007-05-27 2007     05/27
3     23900 2007-05-28 2007     05/28
4    121050 2007-05-29 2007     05/29
5     87500 2007-05-30 2007     05/30
6     46500 2007-05-31 2007     05/31

• step3

library(ggplot2)

qplot(date_1,sumamount,data=lend_club1,geom="line") # 天天貸款金額的時序圖

p<-qplot(month_day,sumamount,data=lend_club1)
p+facet_wrap(~year) #2007-2012 期間每日的貸款金額

library(tidyr)
library(dplyr)
lend_club2<-separate(lend_club1,date_1,c("y","m","d"),sep="-")
head(lend_club2)
  sumamount    y  m  d year month_day
1      2000 2007 05 26 2007     05/26
2     47400 2007 05 27 2007     05/27
3     23900 2007 05 28 2007     05/28
4    121050 2007 05 29 2007     05/29
5     87500 2007 05 30 2007     05/30
6     46500 2007 05 31 2007     05/31

lend_club3<-unite(lend_club2,"y_m",y,m,sep="-",remove = F)
head(lend_club3)
  sumamount     y_m    y  m  d year month_day
1      2000 2007-05 2007 05 26 2007     05/26
2     47400 2007-05 2007 05 27 2007     05/27
3     23900 2007-05 2007 05 28 2007     05/28
4    121050 2007-05 2007 05 29 2007     05/29
5     87500 2007-05 2007 05 30 2007     05/30
6     46500 2007-05 2007 05 31 2007     05/31

qplot(m,sumamount,data=lend_club3,geom=c("boxplot")+facet_wrap(~year) #2007~2012年每個月貸款金額的箱線圖

lend_club4<- lend_club3%>%
  group_by(m,y)%>%
  summarise(total_m=sum(sumamount))

lend_club4
head(lend_club4)
Source: local data frame [6 x 3]
Groups: m [2]

      m     y   total_m
  (chr) (chr)     (dbl)
1    01  2008  32256329
2    01  2009  28523635
3    01  2010  63082946
4    01  2011 171186425
5    01  2012 297667575
6    02  2008  20596688

折線圖 分面
p<-qplot(m,total_m,data=lend_club4)+geom_smooth(aes(group=y,colour=y),method = "lm")

折線圖 分面

p<-qplot(m,total_m,data=lend_club4)+geom_smooth(aes(group=y,colour=y))

p+facet_wrap(~y)

lend<-read.csv("C:\\Users\\liyi\\Desktop\\lend_club.csv")
lend1<-read.csv("C:\\Users\\liyi\\Desktop\\lend_club.csv",header = F)
lend1<-lend1[-1,]
head(lend1)
lend1<-lend1[,c(1,3,9)]
myvar<-c("amount","year","employment")
names(lend1)<-myvar
head(lend1)
str(lend1)
lend1$amountnew<-as.numeric(as.character(lend1$amount))

library(sqldf)

lend2<-sqldf('select sum(V1),V3,V9
             from lend1
             group by V3,V9')
q<-qplot(employment,amountnew,data = lend1,geom=c("boxplot"),colour=lend1$employment)+facet_wrap(~year)
q<- q+theme(axis.text.x=element_text(angle=90,hjust=1,colour="black"),legend.position='none')
q<- q+scale_y_continuous(limits = c(0, 100000))
q