German Credit Data Exploration_2
Dr. Prashant Mishra
3/26/2018
1. Import the Clean data
We already have a clean file “german_credit_full.csv” to import.
credit_dataset <- read.csv("german_credit_full.csv",stringsAsFactors = TRUE)
str(credit_dataset)## 'data.frame': 1000 obs. of 21 variables:
## $ Class : Factor w/ 2 levels "Bad","Good": 2 1 2 2 1 2 2 2 2 1 ...
## $ CheckingAccountStatus : Factor w/ 4 levels "0.to.200","gt.200",..: 3 1 4 3 3 4 4 1 4 1 ...
## $ Duration : int 6 48 12 42 24 36 24 36 12 30 ...
## $ CreditHistory : Factor w/ 5 levels "Critical","Delay",..: 1 4 1 4 2 4 4 4 4 1 ...
## $ Purpose : Factor w/ 10 levels "Business","DomesticAppliance",..: 7 7 3 4 5 3 4 10 7 5 ...
## $ Amount : int 1169 5951 2096 7882 4870 9055 2835 6948 3059 5234 ...
## $ SavingsAccountBonds : Factor w/ 5 levels "100.to.500","500.to.1000",..: 5 4 4 4 4 5 2 4 3 4 ...
## $ EmploymentDuration : Factor w/ 5 levels "0.to.1","1.to.4",..: 4 2 3 3 2 2 4 2 3 5 ...
## $ InstallmentRatePercentage: int 4 2 2 2 3 2 3 2 2 4 ...
## $ Personal : Factor w/ 4 levels "Female.NotSingle",..: 4 1 4 4 4 4 4 4 2 3 ...
## $ OtherDebtorsGuarantors : Factor w/ 3 levels "CoApplicant",..: 3 3 3 2 3 3 3 3 3 3 ...
## $ ResidenceDuration : int 4 2 3 4 4 4 4 2 4 2 ...
## $ Property : Factor w/ 4 levels "CarOther","Insurance",..: 3 3 3 2 4 4 2 1 3 1 ...
## $ Age : int 67 22 49 45 53 35 53 35 61 28 ...
## $ OtherInstallmentPlans : Factor w/ 3 levels "Bank","None",..: 2 2 2 2 2 2 2 2 2 2 ...
## $ Housing : Factor w/ 3 levels "ForFree","Own",..: 2 2 2 1 1 1 2 3 2 2 ...
## $ NumberExistingCredits : int 2 1 1 1 2 1 1 1 1 2 ...
## $ Job : Factor w/ 4 levels "Management.SelfEmp.HighlyQualified",..: 2 2 4 2 2 4 2 1 4 1 ...
## $ NumberPeopleMaintenance : int 1 1 2 2 2 2 1 1 1 1 ...
## $ Telephone : int 1 0 0 0 0 1 0 1 0 0 ...
## $ ForeignWorker : int 1 1 1 1 1 1 1 1 1 1 ...colnames(credit_dataset)## [1] "Class" "CheckingAccountStatus"
## [3] "Duration" "CreditHistory"
## [5] "Purpose" "Amount"
## [7] "SavingsAccountBonds" "EmploymentDuration"
## [9] "InstallmentRatePercentage" "Personal"
## [11] "OtherDebtorsGuarantors" "ResidenceDuration"
## [13] "Property" "Age"
## [15] "OtherInstallmentPlans" "Housing"
## [17] "NumberExistingCredits" "Job"
## [19] "NumberPeopleMaintenance" "Telephone"
## [21] "ForeignWorker"2. Explore Class vs Checking Account Status, Credit History and Employment Duration
library(gmodels)
CrossTable(credit_dataset$CheckingAccountStatus, credit_dataset$Class)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | credit_dataset$Class
## credit_dataset$CheckingAccountStatus | Bad | Good | Row Total |
## -------------------------------------|-----------|-----------|-----------|
## 0.to.200 | 105 | 164 | 269 |
## | 7.317 | 3.136 | |
## | 0.390 | 0.610 | 0.269 |
## | 0.350 | 0.234 | |
## | 0.105 | 0.164 | |
## -------------------------------------|-----------|-----------|-----------|
## gt.200 | 14 | 49 | 63 |
## | 1.270 | 0.544 | |
## | 0.222 | 0.778 | 0.063 |
## | 0.047 | 0.070 | |
## | 0.014 | 0.049 | |
## -------------------------------------|-----------|-----------|-----------|
## lt.0 | 135 | 139 | 274 |
## | 33.915 | 14.535 | |
## | 0.493 | 0.507 | 0.274 |
## | 0.450 | 0.199 | |
## | 0.135 | 0.139 | |
## -------------------------------------|-----------|-----------|-----------|
## none | 46 | 348 | 394 |
## | 44.102 | 18.901 | |
## | 0.117 | 0.883 | 0.394 |
## | 0.153 | 0.497 | |
## | 0.046 | 0.348 | |
## -------------------------------------|-----------|-----------|-----------|
## Column Total | 300 | 700 | 1000 |
## | 0.300 | 0.700 | |
## -------------------------------------|-----------|-----------|-----------|
##
## library(ggplot2)
pl1 = ggplot(credit_dataset, aes(x = CheckingAccountStatus, fill = Class));
pl2 = pl1 + geom_bar()
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl3 = pl1 + geom_bar(position = "fill") + ylab("proportion")
pl3+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
Two points:
- It seems people who have Checking Account but only get amount between 0-200 DM are more likely to be a bad loan profile.
- People who don’t have a checking account might have account else where (other banks).
pl1 = ggplot(credit_dataset, aes(x = CheckingAccountStatus, fill = EmploymentDuration));
pl2 = pl1 + geom_bar()+facet_grid(~Class)
pl2+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl2 = pl1 + geom_bar(position = "fill") + ylab("proportion")+facet_grid(~Class)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = CheckingAccountStatus, fill = CreditHistory));
pl2 = pl1 + geom_bar()+facet_grid(~Class)
pl2+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl2 = pl1 + geom_bar(position = "fill") + ylab("proportion")+facet_grid(~Class)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
2. Explore Class vs Purpose, Personal and Property
CrossTable(credit_dataset$Purpose,credit_dataset$Class)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | credit_dataset$Class
## credit_dataset$Purpose | Bad | Good | Row Total |
## -----------------------|-----------|-----------|-----------|
## Business | 34 | 63 | 97 |
## | 0.825 | 0.354 | |
## | 0.351 | 0.649 | 0.097 |
## | 0.113 | 0.090 | |
## | 0.034 | 0.063 | |
## -----------------------|-----------|-----------|-----------|
## DomesticAppliance | 4 | 8 | 12 |
## | 0.044 | 0.019 | |
## | 0.333 | 0.667 | 0.012 |
## | 0.013 | 0.011 | |
## | 0.004 | 0.008 | |
## -----------------------|-----------|-----------|-----------|
## Education | 22 | 28 | 50 |
## | 3.267 | 1.400 | |
## | 0.440 | 0.560 | 0.050 |
## | 0.073 | 0.040 | |
## | 0.022 | 0.028 | |
## -----------------------|-----------|-----------|-----------|
## Furniture.Equipment | 58 | 123 | 181 |
## | 0.252 | 0.108 | |
## | 0.320 | 0.680 | 0.181 |
## | 0.193 | 0.176 | |
## | 0.058 | 0.123 | |
## -----------------------|-----------|-----------|-----------|
## NewCar | 89 | 145 | 234 |
## | 5.035 | 2.158 | |
## | 0.380 | 0.620 | 0.234 |
## | 0.297 | 0.207 | |
## | 0.089 | 0.145 | |
## -----------------------|-----------|-----------|-----------|
## Others | 5 | 7 | 12 |
## | 0.544 | 0.233 | |
## | 0.417 | 0.583 | 0.012 |
## | 0.017 | 0.010 | |
## | 0.005 | 0.007 | |
## -----------------------|-----------|-----------|-----------|
## Radio.Television | 62 | 218 | 280 |
## | 5.762 | 2.469 | |
## | 0.221 | 0.779 | 0.280 |
## | 0.207 | 0.311 | |
## | 0.062 | 0.218 | |
## -----------------------|-----------|-----------|-----------|
## Repairs | 8 | 14 | 22 |
## | 0.297 | 0.127 | |
## | 0.364 | 0.636 | 0.022 |
## | 0.027 | 0.020 | |
## | 0.008 | 0.014 | |
## -----------------------|-----------|-----------|-----------|
## Retraining | 1 | 8 | 9 |
## | 1.070 | 0.459 | |
## | 0.111 | 0.889 | 0.009 |
## | 0.003 | 0.011 | |
## | 0.001 | 0.008 | |
## -----------------------|-----------|-----------|-----------|
## UsedCar | 17 | 86 | 103 |
## | 6.253 | 2.680 | |
## | 0.165 | 0.835 | 0.103 |
## | 0.057 | 0.123 | |
## | 0.017 | 0.086 | |
## -----------------------|-----------|-----------|-----------|
## Column Total | 300 | 700 | 1000 |
## | 0.300 | 0.700 | |
## -----------------------|-----------|-----------|-----------|
##
## CrossTable(credit_dataset$Personal,credit_dataset$Class)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | credit_dataset$Class
## credit_dataset$Personal | Bad | Good | Row Total |
## ------------------------|-----------|-----------|-----------|
## Female.NotSingle | 109 | 201 | 310 |
## | 2.753 | 1.180 | |
## | 0.352 | 0.648 | 0.310 |
## | 0.363 | 0.287 | |
## | 0.109 | 0.201 | |
## ------------------------|-----------|-----------|-----------|
## Male.Divorced.Seperated | 20 | 30 | 50 |
## | 1.667 | 0.714 | |
## | 0.400 | 0.600 | 0.050 |
## | 0.067 | 0.043 | |
## | 0.020 | 0.030 | |
## ------------------------|-----------|-----------|-----------|
## Male.Married.Widowed | 25 | 67 | 92 |
## | 0.245 | 0.105 | |
## | 0.272 | 0.728 | 0.092 |
## | 0.083 | 0.096 | |
## | 0.025 | 0.067 | |
## ------------------------|-----------|-----------|-----------|
## Male.Single | 146 | 402 | 548 |
## | 2.059 | 0.883 | |
## | 0.266 | 0.734 | 0.548 |
## | 0.487 | 0.574 | |
## | 0.146 | 0.402 | |
## ------------------------|-----------|-----------|-----------|
## Column Total | 300 | 700 | 1000 |
## | 0.300 | 0.700 | |
## ------------------------|-----------|-----------|-----------|
##
## CrossTable(credit_dataset$Property,credit_dataset$Class)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | credit_dataset$Class
## credit_dataset$Property | Bad | Good | Row Total |
## ------------------------|-----------|-----------|-----------|
## CarOther | 102 | 230 | 332 |
## | 0.058 | 0.025 | |
## | 0.307 | 0.693 | 0.332 |
## | 0.340 | 0.329 | |
## | 0.102 | 0.230 | |
## ------------------------|-----------|-----------|-----------|
## Insurance | 71 | 161 | 232 |
## | 0.028 | 0.012 | |
## | 0.306 | 0.694 | 0.232 |
## | 0.237 | 0.230 | |
## | 0.071 | 0.161 | |
## ------------------------|-----------|-----------|-----------|
## RealEstate | 60 | 222 | 282 |
## | 7.153 | 3.066 | |
## | 0.213 | 0.787 | 0.282 |
## | 0.200 | 0.317 | |
## | 0.060 | 0.222 | |
## ------------------------|-----------|-----------|-----------|
## Unknown | 67 | 87 | 154 |
## | 9.365 | 4.013 | |
## | 0.435 | 0.565 | 0.154 |
## | 0.223 | 0.124 | |
## | 0.067 | 0.087 | |
## ------------------------|-----------|-----------|-----------|
## Column Total | 300 | 700 | 1000 |
## | 0.300 | 0.700 | |
## ------------------------|-----------|-----------|-----------|
##
## pl1 = ggplot(credit_dataset, aes(x = Purpose, fill = Class));
pl1 + geom_bar()+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 + geom_bar(position = "fill") + ylab("proportion")+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Purpose, fill = Personal));
pl2 = pl1 + geom_bar()+facet_grid(~Class)
pl2+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl2 = pl1 + geom_bar(position = "fill") +ylab("proportion")+facet_grid(~Class)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Purpose, fill = Property));
pl2 = pl1 + geom_bar()+facet_grid(~Class)
pl2+ theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl2 = pl1 + geom_bar(position = "fill") + ylab("proportion")+facet_grid(~Class)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
New Car loan seems to be worse than used car loans and any other loan.
Property doesn’t seem to play a huge role as a deciding factor.
pl1 = ggplot(credit_dataset, aes(x = Purpose, y = Amount, color=Purpose));
pl2 = pl1 + geom_boxplot()+facet_grid(~Class)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Personal, y = Amount, color=Personal));
pl2 = pl1 + geom_boxplot()+facet_grid(~Class)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Job, y = Amount, color=Job));
pl2 = pl1 + geom_boxplot()+facet_wrap(~Class,ncol = 2)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Housing, y = Amount, color=Class));
pl2 = pl1 + geom_boxplot()+facet_wrap(~Personal,ncol=2)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Purpose, y = Duration, color=Class));
pl2 = pl1 + geom_boxplot()+facet_wrap(~Personal,ncol=2)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Personal, y = Amount, color=Class));
pl2 = pl1 + geom_boxplot()+facet_wrap(~Purpose, ncol = 5)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
- Used car data show that that Female.NotSingle tend to take higher loan for Used car, and are more likely bad loans.
- Unsual data for Others, we might need to get more data on that.
pl1 = ggplot(credit_dataset, aes(x = Personal, y = Amount, color=Class));
pl2 = pl1 + geom_boxplot()+facet_wrap(~Job,ncol=2)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
1. Male single, Unemployed and Unskilled— high risk 2. Female Not Single — Also high risk 3. Male Divored/Separated Unskilled Resident – high risk
pl1 = ggplot(credit_dataset, aes(x = Personal, y = Amount, color=Class));
pl2 = pl1 + geom_boxplot()+facet_wrap(~CheckingAccountStatus,ncol=2)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
pl1 = ggplot(credit_dataset, aes(x = Personal, y = Amount, color=Class));
pl2 = pl1 + geom_boxplot()+facet_wrap(~CreditHistory,ncol=3)
pl2 + theme(axis.text.x = element_text(angle = 90, hjust = 1))
Just some detective skill: We can see the outlier in NoCredit.AllPaid, Female.NotSingle. Some one from Highskilled, management or self employed category with checking account status between 0 to 200 took the loan for others category.