R语言25-Prosper 贷款数据分析1_prosperloandata.csv

Prosper 贷款数据分析

数据导入

• 工作目录设置
  首先查看当前工作路径，对其进行对应更改

getwd()
setwd('C:/Users/blabla/Desktop/数据分析/R语言')
getwd()
list.dirs()   #罗列目录下所有文件夹
list.files()   #罗列目录下所有文件
1
2
3
4
5

• 导入数据

pf <- read.csv('C:/Users/孔啊吱/Desktop/数据分析/R语言/prosperLoanData.csv')
1

• 查看数据

names(pf)    #查看字段名
str(pf)      #The str() function gives us the variable names and their types.
dim(pf)      #查看数据大小
head(pf,2)   #查看前两行
1
2
3
4

变量解释：

• 划分数据集

data <- pf[c("ListingKey","LoanOriginationDate",
             "BorrowerRate","Occupation","Term",
             "EmploymentStatus","BorrowerState",
             "ListingCategory..numeric.","IncomeRange",    #注意..numeric.
             "CreditGrade","ProsperRating..Alpha.","LenderYield",
             "LoanStatus","EmploymentStatusDuration",
             "IsBorrowerHomeowner","CreditScoreRangeLower",
             "CreditScoreRangeUpper","InquiriesLast6Months",
             "DelinquenciesLast7Years","BankcardUtilization",
             "DebtToIncomeRatio","StatedMonthlyIncome",
             "LoanOriginalAmount")]
1
2
3
4
5
6
7
8
9
10
11

• 初步数据整理(数据预处理)

1.借款人信用评级(CreditGrade/ProsperRating(Alpha))：CreditGrade为2009年7月1日前的数据，ProsperRating(Alpha)为2009年7月1日后的数据，将其进行合并

temp <- pf[c("CreditGrade","ProsperRating..Alpha.")]

for (i in 1:nrow(temp)){
  if (temp[i,1] == ""){
    temp$new[i] <- temp[i,2]
  }else{
    temp$new[i] <- temp[i,1]
  }
}

temp$new <- factor(temp$new,
                   levels = c(2,3,4,5,6,7,8),
                   labels = c("A","AA","B","C",
                              "D","E","HR"))
data$creditlevel <- temp$new
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

data$Phase[as.character(data$LoanOriginationDate) > "2009-07-01"] <- "After 2009"    #标记生成年份
data$Phase[as.character(data$LoanOriginationDate) < "2009-07-01"] <- "Before 2009"
table(data$Phase)

After 2009 Before 2009 
      84997       28940
1
2
3
4
5
6

2.信用评分(CreditScoreRangeLower/CreditScoreRangeUpper):使用消费者信用评级机构提供的借款人信用评分范围下限值/上限值的均值作为最终评估的信用评分。

data$creditscore <- (data$CreditScoreRangeLower + 
  data$CreditScoreRangeUpper)/2

head(data$creditscore,2)
1
2
3
4

3.贷款状态处理：平台把贷款状态分为7大种:Cancelled(取消)、Chargedoff(冲销，投资人有损失)、Completed(正常完成，投资人无损失)、Current(贷款还款中)、Defaulted(坏账，投资人有损失)、FinalPaymentInProgress(最后还款中，投资人无损失)、PastDue(逾期还款，投资人无损失)。为了方便统计及后续的预测模型，我们将贷款状态分为两类：问题贷款和正常贷款。正常贷款包括：Completed、Current、FinalPaymentInProgress三种，其余都归为问题贷款。(资料来自https://zhuanlan.zhihu.com/p/39812067)

#levels(data$LoanStatus) 
table(data$LoanStatus) 
1
2

#将Completed、Current、FinalPaymentInProgress三种划分为正常贷款
data$newLoanStatus[data$LoanStatus == "Current"] <- "normalloan"
data$newLoanStatus[data$LoanStatus == "FinalPaymentInProgress"] <- "normalloan"
data$newLoanStatus[data$LoanStatus == "Completed"] <- "normalloan"

#将所有逾期的贷款均归为Past Due
PastDue <- c("Past Due (1-15 days)","Past Due (16-30 days)",
             "Past Due (31-60 days)","Past Due (61-90 days)",
             "Past Due (91-120 days)","Past Due (>120 days)")
data$LoanStatus[data$LoanStatus %in% PastDue] <- "PastDue"   #attention
#将其余均划分为问题贷款(problemloan)
problemloan <- c("PastDue","Cancelled","Chargedoff","Defaulted")
data$newLoanStatus[data$LoanStatus %in% problemloan] <- "problemloan"#attention
1
2
3
4
5
6
7
8
9
10
11
12
13

4.缺失值处理
4.1缺失值查看

sapply(data,function(x)sum(is.na(x)))
1

                ListingKey       LoanOriginationDate              BorrowerRate 
                         0                         0                         0 
                Occupation                      Term          EmploymentStatus 
                         0                         0                         0 
             BorrowerState ListingCategory..numeric.               IncomeRange 
                         0                         0                         0 
               CreditGrade     ProsperRating..Alpha.               LenderYield 
                         0                         0                         0 
                 LoanStatus  EmploymentStatusDuration       IsBorrowerHomeowner 
                       2067                      7625                         0 
      CreditScoreRangeLower     CreditScoreRangeUpper      InquiriesLast6Months 
                       591                       591                       697 
   DelinquenciesLast7Years       BankcardUtilization         DebtToIncomeRatio 
                       990                      7604                      8554 
       StatedMonthlyIncome        LoanOriginalAmount               creditlevel 
                         0                         0                       272 
                     Phase               creditscore             newLoanStatus 
                         0                       591                      2067
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

4.2缺失值处理(使用均值替代)

data[is.na(data$DebtToIncomeRatio),"DebtToIncomeRatio"] <- mean(data$DebtToIncomeRatio,na.rm = TRUE)

data[is.na(data$EmploymentStatusDuration),"EmploymentStatusDuration"] <- mean(data$EmploymentStatusDuration,na.rm = TRUE)
1
2
3