Trainity Data Analytics Trainee Task 6

Name : ADVAIT GURUNATH CHAVAN
Email : advaitchavan135@gmail.com
Task 6: Bank Loan Case Study (Final Project - 2),
Tech Stack Used: Microsoft Excel
Analysis done on the following points:
To identify patterns which indicate if a client has difficulty paying their

installments which may be used for taking actions such as denying the
loan, reducing the amount of loan, lending (to risky applicants) at a
higher interest rate, etc. This will ensure that the consumers capable of
repaying the loan are not rejected.
Analysis is being done into two parts or say two dataset wiz:
1. Application data
2. Previous application data
The cleaned and analyzed data in the form of excel sheets have
been uploaded to Google Drive also the excel sheets are large files
due to vastness of data, so they won’t be visible on google excel
sheets online they need to be downloaded and seen offline using
Microsoft Excel 2019
Application Dataset – NULL values
Firstly the percentage of null values needs to be analyzed and those columns that
have more than 50% of the null data have to be dropped
And those columns with less than 50% of the null data have to be replaced with
mean or median or the highest occurring categorical variables
ALL THE COLUMN NAME WHICH ARE HIGHLIGHTED IN BLUE NEED TO BE DROPPED DOWN
AS THEY HAVE NULL VALUES GREATER THAN OR EQUAL TO 50%
Column name Total number of null values Percentage of null value in that column ROUND PER
OWN_CAR_AGE 202930 65.99113528 66
EXT_SOURCE_1 173379 56.38139774 56
APARTMENTS_AVG 156061 50.74972928 51
BASEMENTAREA_AVG 179943 58.51595553 59
YEARS_BUILD_AVG 204488 66.49778382 66
COMMON_AREA_AVG 214865 69.87229725 70
ELEVATORS_AVG 163891 53.29597966 53
ENTRANCES_AVG 154828 49.70488861 50
FLOORSMAX_AVG 153021 49.76114676 50
FLOORSMIN_AVG 208642 67.84862981 68
LANDAREA_AVG 182590 59.37673774 59
LIVINGAPARTMENTS_AVG 210199 68.35495316 68
LIVINGAREA_AVG 154350 50.19332642 50
NONLIVINGAPARTMENTS_AVG 213514 69.43296337 69
NONLIVINGAREA_AVG 169682 55.17916432 55
APARTMENTS_MODE 156061 50.74972928 51
BASEMENTAREA_MODE 179943 58.51595553 59
YEARS_BUILD_MODE 204488 66.49778382 66
COMMON_AREA_MODE 214865 69.87229725 70
ELEVATORS_MODE 163891 53.29597966 53
ENTRANCES_MODE 154828 50.34876801 50
FLOORSMAX_MODE 153020 49.76082156 50
FLOORSMIN_MODE 208642 67.84862981 68
LANDAREA_MODE 182590 59.37673774 59
LIVINGAPARTMENTS_MODE 210199 68.35495316 68
LIVINGAREA_MODE 154350 50.19332642 50
NONLIVINGAPARTMENTS_MODE 213514 69.43296337 69
NONLIVINGAREA_MODE 169682 55.17916432 55
APARTMENTS_MEDIAN 156061 50.74972928 51
BASEMENTAREA_MEDIAN 179943 58.51595553 59
YEARS_BUILD_MEDIAN 204488 66.49778382 66
COMMON_AREA_MEDIAN 214865 69.87229725 70
ELEVATORS_MEDIAN 163891 53.29597966 53
ENTRANCES_MEDIAN 154828 50.34876801 50
FLOORSMAX_MEDIAN 153020 49.76082156 50
FLOORSMIN_MEDIAN 208642 67.84862981 68
LANDAREA_MEDIAN 182590 59.37673774 59
LIVINGAPARTMENTS_MEDIAN 210199 68.35495316 68
LIVINGAREA_MEDIAN 154350 50.19332642 50
NONLIVINGAPARTMENTS_MEDIAN 213514 69.43296337 69
NONLIVINGAREA_MEDIAN 169682 55.17916432 55
FONDKAPREMONT_MODE 210295 68.38617155 68
HOUSETYPE_MODE 154297 50.17609126 50
WALLSMATERIAL_MODE 156341 50.84078293 51
Application
ApplicationDataset
Dataset– NULL values
ALL THE COLUMN NAME WHICH ARE HIGHLIGHTED IN GREEN NEED TO BE DROPPED DOWN
AS THEY ARE IRRELEVANT COLUMNS FOR DOING OUR ANALYSIS
Column name Total number of null values Percentage of null value in that column ROUND PER
FLAG_MOBIL 1 0.000325192 0
FLAG_EMPLOY_PHONE 55387 18.01138821 18
FLAG_WORK_PHONE 0 0 0
FLAG_CONT_MOBILE 0 0 0
FLAG_PHONE 0 0 0
FLAG_EMAIL 0 0 0
CNT_FAMILY_MEMBERS 2 0.000650383 0
REGION_RATING_CLENT 0 0 0
REGION_RATING_CLENT_W_CITY 0 0 0
EXT_SOURCE_3 60965 19.82530706 20
YEAR_BEGINEXPLUATATION_AVG 150008 48.78134441 49
YEAR_BEGINEXPLUATATION_MODE 150007 48.78101922 49
YEAR_BEGINEXPLUATATION_MEDIAN 150007 48.78101922 49
TOTAL_AREA_MODE 148431 48.26851722 48
EMERGENCYSTATE_MODE 145755 47.39830445 47
DAYS_LAST_PHONE_CHANGE 1 0.000325192 0
FLAG DOC 2 0 0 0
FLAG DOC 3 0 0 0
FLAG DOC 4 0 0 0
FLAG DOC 5 0 0 0
FLAG DOC 6 0 0 0
FLAG DOC 7 0 0 0
FLAG DOC 8 0 0 0
FLAG DOC 9 0 0 0
FLAG DOC 10 0 0 0
FLAG DOC 11 0 0 0
FLAG DOC 12 0 0 0
FLAG DOC 13 0 0 0
FLAG DOC 14 0 0 0
FLAG DOC 15 0 0 0
FLAG DOC 16 0 0 0
FLAG DOC 17 0 0 0
FLAG DOC 18 0 0 0
FLAG DOC 19 0 0 0
FLAG DOC 20 0 0 0
FLAG DOC 21 0 0 0
Application
ApplicationDataset
Replacing Blanks in Occupation_Type column of the Application

Dataset with the highest occurring categorical variable
Row Labels Count of OCCUPATION_TYPE
Total count of each Occupation_type Accountants

Cleaning staff
9813
4653
Cooking staff 5946
Core staff 27570
Drivers 18603
55186 High skill tech staff 11380
HR staff 563
IT staff 526
Laborers 55186
Low-skill Laborers 2093
Managers 21371
Medicine staff 8537
Private service staff 2652

32102 Realty agents 751
Sales staff 32102
27570 Secretaries 1305
Security staff 6721
Total
21371
18603 Waiters/barmen staff 1348
(blank)
Grand Total 211120
11380
9813
8537
6721
4653
5946
563 526 2093 2652 751 1305 1348 Highest occurring
categorical variable is
‘Laborers’
Application
ApplicationDataset
Replacing Blanks in AMT_ANNUTIY column of the Application Dataset

with the median of the AMT_ANNUITY as there exists outliers in the
AMT_ANNUITY column
Median of AMT_ANNUITY
24903
Replacing Blanks with Median

Application
ApplicationDataset
Replacing Blanks in AMT_GOODS_PRICE column of the Application

Dataset with the median of the AMT_GOODS_PRICE as there exists
outliers in the AMT_GOODS_PRICE column
Median of AMT_GOODS_PRICE
450000
Replacing Blanks with Median

Application
ApplicationDataset
Replacing Blanks in Name_Type_Suite column of the Application

Dataset with the highest occurring categorical variable
Row Labels Count of NAME_TYPE_SUITE
Total count of each Name_Type_Suite Children 3267
Family 40149
Group of people 271
Other_A 866
Other_B 1770
248526 Spouse, partner 11370
Unaccompanied 248526
(blank)
Grand Total 306219
Highest occurring
Total
‘Unaccompanied’
40149
3267 271 866 1770 11370
Application
ApplicationDataset
Dataset– NULL values Row Labels
Advertising
Count of ORGANIZATION_TYPE
429
Agriculture 2454
Bank 2507
Replacing Blanks in Organization_type column of the Application

Business Entity Type 1 5984
Dataset with the highest occurring categorical variable Cleaning

Construction
260
6721
Culture 379
Electricity 950
Total count of each Organization_type Highest occurring Emergency

Government
560
10404
Hotel 966
Housing 2958
Industry: type 1 1039
‘Business Entity Type 3’ Industry: type 10

Industry: type 11
109
2704
67992 Industry: type 13 67
55374 Industry: type 6 112
Industry: type 8 24
Insurance 597
Kindergarten 6880
Legal Services 305
Medicine 11193
38412 Military 2634
Mobile 317
Other 16683
Police 2341
Total Postal 2157
Realtor 396
Religion 85
Restaurant 1811
School 8893
16683 Security 3247
Security Ministries 1974
10553 10404 11193 Self-employed 38412

8893 7831
Services 1575
429
2454 5984
2507 6721
260 379950560 966
2958
1039
109
2704
36967458
3278
877599112
1307243368 6880
597 305 2634
317 2341
2157
396851811 3247
1974 1575
577348
1900
349264 49631 2012204
1187
5398
1327
Telecom 577
Trade: type 1 348
Trade: type 2 1900
Insurance
School
Hotel
Construction
Religion
Services
Trade: type 4
Industry: type 2
Industry: type 6
Housing
Mobile
Police
Realtor
Telecom
Transport: type 1
Bank
Culture
Other
Security Ministries
Trade: type 1
Trade: type 2
Trade: type 3
Trade: type 5
Trade: type 6
Trade: type 7
XNA
Government
Military
Self-employed
Advertising
Industry: type 1
Industry: type 3
Industry: type 4
Industry: type 5
Industry: type 7
Industry: type 8
Industry: type 9
Postal
Agriculture
Security
Transport: type 2
Transport: type 3
Transport: type 4
Business Entity Type 1
Cleaning
Electricity
Emergency
Kindergarten
Medicine
Restaurant
University
Industry: type 10
Industry: type 11
Industry: type 12
Industry: type 13
Legal Services
Trade: type 3 3492

Trade: type 4 64
Trade: type 5 49
Trade: type 6 631
Trade: type 7 7831
Transport: type 1 201
University 1327
XNA 55374
Grand Total 307511
Application
ApplicationDataset
Google Drive Link for Excel sheet of Analysis of Null values done:-
application_data.xlsx - Google
Drive
Application Dataset – Outliers
First outlier is in AMT_ANNUITY

which is greater than 250000
this outlier is replaced with 24903
median of AMT_ANNUITY
Quartiles at AMT_INCOME_TOTAL
Here we can observe that there is huge difference MIN 25650
between 25% 112500
the 25%, 50% and 75% quartile and this is due to presence 50% 147150
of outliers 75% 202500
But since the amount of total income varies from person to MAX 117000000
person
we will not remove the outliers
outliers at extreme points i.e. max 1.700x10^8
From the chart it is clear that outliers lie in the 98% and near AMT_CREDIT
max side of the box plot Quartiles at AMT_CREDIT
Also there is a significant difference between the 75% MIN 45000
quartile and the max value and this is due the presence of 25% 270000
the outliers 50% 513531
But since the amount of credit varies from person to person 75% 808650
we will not remove the outliers MAX 4050000
DAYS_OF_BIRTH
Quartiles at DAYS_BIRTH
As seen from the boxplot it is clear that there are no MAX 25,229.00
outliers 75% 19,682.00
The data of DAYS_BIRTH is well distributed 50% 15,750.00
25% 12,413.00
MIN 7,489.00
DAYS_EMPLOYED
Quartiles at DAYS_EMPLOYED
There exists only 1 outlier i.e. + or - 365243 MAX 17912.00
75% 2760.00
Replace with median
1213.00 50% 1213.00
25% 289.00
MIN 365243.00
Google Drive Link for Excel sheet of Analysis of Outliers and cleaned Data
done:-
application_data_cleaned.xlsx - Google Drive

Application Dataset – Analysis
TARGET VARIABLE
Count of
Row Labels TARGET
0
1
282686
24825
Target Variable
Grand Total 307511
1
8%
0
1
0
The Target Variable Pie chart 92%
shows that almost 92%
of the total clients had
no problem during payment
while 8% of the clients
had some or the other problem
0 → No payment issues
1 → Had some payment issues
GENDER VARIABLE
Count of
Row Labels
F
CODE_GENDER
202448
CODE_GENDER
M 105059
XNA 4
Grand Total 307511
0%
34% F
M
XNA
66%
From the GENDER_VARIABLE pie chart

we can infer that almost 66% of
the clients are female and 34% of the
clients are Male
The 4 of the appicants have gender as XNA
which can be ignored
NAME_HOUSING_TYPE
Row Labels Count of NAME_HOUSING_TYPE

Co-op apartment 1122
House / apartment 272868
Municipal apartment 11183
Office apartment 2617
Rented apartment 4881
With parents 14840
Grand Total 307511
Total count of each Housing Type
272868
Total
1122 11183 2617 4881 14840

Co-op apartment House / apartment Municipal apartment Office apartment Rented apartment With parents
Total 1122 272868 11183 2617 4881 14840
NAME_HOUSING_TYPE
Percentage of
Row Labels NAME_HOUSING_TYPE
Co-op apartment 0.36%
House / apartment 88.73%
Municipal apartment 3.64%
Office apartment 0.85%
Rented apartment 1.59%
From the bar graphs of count and percentage
With parents 4.83%
Grand Total 100.00% The bank can target those groups who do not have
their
own apartment i.e. the bank may consider the

Total percentage of each Housing Type people
living in Co-op apartment, Municipal Apartment,

Rented
88.73%
Apartment and people living with their parents
Total
0.36% 3.64% 0.85% 1.59% 4.83%

Co-op House / Municipal Office Rented
With parents
apartment apartment apartment apartment apartment
Total 0.36% 88.73% 3.64% 0.85% 1.59% 4.83%
Univariate Analysis
AGE GROUP
Count of
Row Labels YEARS_BIRTH_RANGE Total count of each age group of the banks
20-30 48869
31-40 82770 applicants
41-50 75509 90000
51-60 67955
61-70 32408 80000 82770
Grand Total 307511
75509
70000
67955
60000
From the adjacent bar plot we can
50000
infer that most of the applicants 48869
Total
belong to the Age Group ’31-40’ 40000
30000 32408
20000
10000
0
20-30 31-40 41-50 51-60 61-70
Univariate Analysis
AGE GROUP
Count of TARGET Column Labels
Row Labels 0 Grand Total
20-30 43276 43276
31-40 74961 74961
41-50 69784 69784
51-60 63853 63853
61-70 30812 30812
Grand Total 282686 282686
Clients Age Group with no Payment issues From the adjacent Bar plot we
can infer that
74961
69784
clients/applicants in the Age
63853 Group ’31-40’ are having the
highest number when it
43276 0
comes to doing/returning
Payment to Banks
30812
20-30 31-40 41-50 51-60 61-70

Univariate Analysis
AGE GROUP
20-30 5593 5593
31-40 7809 7809
41-50 5725 5725
51-60 4102 4102
61-70 1596 1596
Clients Age Group with payment issues

From the adjacent Bar plot we
can infer that
7809
clients/applicants in the Age
Group ’31-40’ are having the
5593 5725
highest number of payment
4102
1
issues when it comes to
doing/returning Payment to
Banks
1596
20-30 31-40 41-50 51-60 61-70

Univariate Analysis
Client amount credit range

HIGH 93297 93297
LOW 98720 98720
MEDIUM 90669 90669
Client amount credit range without payment

issues
From the adjacent Bar plot we can
98720 infer that clients belonging to ‘Low’
income range have the highest count
when it comes to clients with no
HIGH
93297 LOW
payment issues
MEDIUM
90669
HIGH LOW MEDIUM

Univariate Analysis
Client amount credit range

HIGH 6600 6600
LOW 8442 8442
MEDIUM 9783 9783
Client amout credit range with payment issue
From the adjacent Bar plot we can

infer that clients belonging to
9783
‘Medium’ income range have the
8442
highest count when it comes to
6600
clients with payment issues
1
HIGH LOW MEDIUM

Univariate Analysis
OCCUPATION_TYPE
Grand Clients occupation type with no payment issues
Row Labels 0 Total
Accountants 9339 9339
Cleaning staff 4206 4206 139461
Cooking staff 5325 5325
Core staff 25832 25832
Drivers 16496 16496
High skill tech staff 10679 10679
HR staff 527 527
IT staff 492 492
Laborers 139461 139461
Low-skill Laborers 1734 1734
Managers 20043 20043 0
Medicine staff 7965 7965
Private service staff 2477 2477 25832 29010
Realty agents 692 692 20043
16496
9339 10679
Sales staff 29010 29010 4206 5325 7965 5999
527 492 1734 2477 692 1213 1196
Secretaries 1213 1213
Security staff 5999 5999
Waiters/barmen staff 1196 1196
From the above bar plot we can infer that clients with occupation_type ‘Laborers’ have the highest
number of count when it comes to clients with no payment issues
Univariate Analysis
OCCUPATION_TYPE
Grand Clients occupation_type with payment issues
Row Labels 1 Total
Accountants 474 474
Cooking staff 621 621
Core staff 1738 1738
Drivers 2107 2107
High skill tech staff 701 701
HR staff 36 36
IT staff 34 34
Laborers 12116 12116
Low-skill Laborers 359 359
Managers 1328 1328 1
3092
Medicine staff 572 572
2107
Private service staff 175 175 1738
1328
Realty agents 59 59 474 447 621 701
359 572 722
36 34 175 59 92 152
Sales staff 3092 3092
Secretaries 92 92
Security staff 722 722
Waiters/barmen staff 152 152
From the above bar plot we can infer that clients with occupation_type ‘Laborers’ have the highest
number of count when it comes to clients with payment issues
Univariate Analysis
NAME_INCOME_TYPE
Grand NAME_INCOME_TYPE with no payment issues
Row Labels 0 Total
Businessman 10 10 143550
Commercial associate 66257 66257
Maternity leave 3 3
Pensioner 52380 52380
State servant 20454 20454
Student 18 18
Unemployed 14 14
Working 143550 143550
66257 0
52380
20454
10 3 18 14
BUSINESSMAN COMMERCIAL MATERNITY PENSIONER STATE SERVANT STUDENT UNEMPLOYED WORKING

ASSOCIATE LEAVE
From the above Bar plot we can infer that clients having income_type as ‘WORKING’ have the
highest count when it comes to clients with no payment issues
Univariate Analysis
NAME_INCOME_TYPE
Grand NAME_TYPE_INCOME with payment issues
Row Labels 1 Total
15224
Commercial associate 5360 5360
Maternity leave 2 2
Pensioner 2982 2982
State servant 1249 1249
Unemployed 8 8
Working 15224 15224
1
5360
2982
1249
2 8
COMMERCIAL MATERNITY LEAVE PENSIONER STATE SERVANT UNEMPLOYED WORKING

ASSOCIATE
From the above Bar plot we can infer that clients having income_type as ‘WORKING’ have the
highest count when it comes to clients with payment issues
Univariate Analysis
AMT_TOTAL INCOME
Row Labels 0
Grand
Total
AMT_TOTAL_INCOME with no payment
HIGH 7595 7595 issues
LOW 201045 201045
MEDIUM 74046 74046
201045
74046
7595
HIGH LOW MEDIUM
From the above Bar plot we can infer that client having the total income range as ‘LOW’
have the highest count when it comes to clients having no payment issues
Univariate Analysis
AMT_TOTAL INCOME
AMT_TOTAL_INCOME with payment
Grand
Row Labels 1 Total issues
HIGH 468 468 18551
LOW 18551 18551
MEDIUM 5806 5806
5806
468
HIGH LOW MEDIUM
From the above Bar plot we can infer that client having the total income range as ‘LOW’
have the highest count when it comes to clients having payment issues
Univariate Analysis
CNT_FAMILY_MEMBERS
Count of CNT_FAM_MEMBERS Column Labels
Row Labels 0
Grand
Total
CNT_FAMILY_MEMBERS with no payment
1
2
62172
146350
62172
146350
issues
3 47993 47993
4 22561 22561
146350
5 3151 3151
6 353 353
7 75 75
8 14 14
9 6 6 0
10 2 2 62172
12 2 2 47993
14 2 2
15 1 1 22561 3151 353 75 14 6 2 2 2 1 2 2
16 2 2
1 2 3 4 5 6 7 8 9 10 12 14 15 16 20
20 2 2
From the above Bar plot we can infer that clients having total count of family
members as 2 have the highest count when it comes to clients having no payment
issues
Univariate Analysis
CNT_FAMILY_MEMBERS
Count of CNT_FAM_MEMBERS Column Labels
Row Labels 1
Grand
Total
CNT_FAMILY_MEMBERS with payment issues
1 5675 5675
2 12009 12009 12009
3 4608 4608
4 2136 2136
5 327 327
6 55 55
7 6 6
8 6 6 5675 1
10 1 1 4608
11 1 1
13 1 1
2136
327 55 6 6 1 1 1
1 2 3 4 5 6 7 8 10 11 13
From the above Bar plot we can infer that clients having total count of family
members as 2 have the highest count when it comes to clients having payment
issues
Univariate Analysis for TARGET variable
CODE_GENDER
Count of CODE_GENDER Column Labels

Grand
Row Labels 0 1 Total 188278
F 188278 14170 202448
M 94404 10655 105059
XNA 4 4
Grand Total 282686 24825 307511
0
94404 1
14170 10655 4
F M XNA
From the above Bar Plot we can infer that Clients with CODE_GENDER = ‘F’ have the highest
number of non-defaulters i.e. 188278-14170 = 174108
NAME_INCOME_TYPE
Count of NAME_INCOME_TYPE Column Labels

Grand
Businessman 10 10
Commercial associate 66257 5360 71617
Maternity leave 3 2 5
Pensioner 52380 2982 55362
State servant 20454 1249 21703
Student 18 18
Unemployed 14 8 22 66257
Working 143550 15224 158774 52380 0
Grand Total 282686 24825 307511 1
10 5360 3 2 2982 20454

1249 18 14 8 15224
From the adjacent Bar Plot we can infer

that clients having NAME_INCOME_TYPE =
‘WORKING’ having the highest count of
Non-defaulters i.e.
143550-15224 = 128326
NAME_EDUCATION_TYPE
Count of
NAME_EDUCATION_TYPE Column Labels
Grand
Row Labels 0 1 Total
198867
Academic degree 161 3 164
Higher education 70854 4009 74863
Incomplete higher 9405 872 10277
Lower secondary 3399 417 3816
Secondary / secondary 0
special 198867 19524 218391 1
70854
Grand Total 282686 24825 307511
161 3 4009 9405 872 3399 417 19524
ACADEMIC HIGHER INCOMPLETE LOWER SECONDARY /

DEGREE EDUCATION HIGHER SECONDARY SECONDARY
SPECIAL
From the above Bar Plot we can infer that clients having
NAME_EDUCATION_TYPE = ‘SECONDARY/SECONDARY SPECIAL’ have the highest count for Non-
defaulters i.e.
198867-19524 = 179343
NAME_FAMILY_STATUS
Count of
NAME_FAMILY_STATUS Column Labels
Grand
Civil marriage 26814 2961 29775 181582
Married 181582 14850 196432
Separated 18150 1620 19770
Single / not married 40987 4457 45444
Unknown 2 2
Widow 15151 937 16088
Grand Total 282686 24825 307511
0
1
From the adjacent Bar Plot we can
infer that clients having 40987
NAME_FAMILY_STATUS = ‘MARRIED’ 26814
have the highest count of Non- 2961 14850 18150 1620 4457 2 15151 937
defaulters i.e. CIVIL MARRIED SEPARATED SINGLE / NOT UNKNOWN WIDOW

181582 – 14850 = 166732 MARRIAGE MARRIED
NAME_HOUSING_TYPE
300000
Count of NAME_HOUSING_TYPE Column Labels
251596
Grand
250000
Co-op apartment 1033 89 1122
200000
House / apartment 251596 21272 272868
Municipal apartment 10228 955 11183
Office apartment 2445 172 2617 150000
0
Rented apartment 4280 601 4881 1
With parents 13104 1736 14840 100000
Grand Total 282686 24825 307511
50000
21272
10228 13104
1033 89 955 2445 172 4280 601 1736
0
Co-op House / Municipal Office Rented With parents
apartment apartment apartment apartment apartment
From the above Bar Plot we can infer that clients having NAME_HOUSING_TYPE =
‘House/Apartment’ have the highest count of Non-defaulters i.e.
251596-21272 = 230324
OCCUPATION_TYPE
Count of Column 160000
OCCUPATION_TYPE Labels
Grand 139461
Accountants 9339 474 9813
Cooking staff 5325 621 5946
120000
Core staff 25832 1738 27570
Drivers 16496 2107 18603 100000
High skill tech staff 10679 701 11380
HR staff 527 36 563
IT staff 492 34 526 80000
Laborers 139461 12116 151577
Low-skill Laborers 1734 359 2093 60000 0
Managers 20043 1328 21371
Medicine staff 7965 572 8537 25832 1
1734 20043
Private service staff 2477 175 2652 40000 93394206 5999
16496 29010
Realty agents 692 59 751 5325
Sales staff 29010 3092 32102
10679 527 492 7965 2477 1196
20000 12116 692 1213
Secretaries 1213 92 1305
Security staff 5999 722 6721 474 447 621 1738 2107 701 36 34 359 1328 572 175 59 3092 92 722 152
Waiters/barmen staff 1196 152 1348 0
Grand Total 282686 24825 307511
From the adjacent Bar plot we can infer that clients

having occupation_type = ‘Laborers’ have the highest
count for Non-defaulters i.e.
139461-12116 = 127345
Bivariate Analysis for TARGET variable
Target 0: Total_income_range vs Code_gender

Grand Total_income_range vs Code_Gender
Row Labels 0 Total 160000
HIGH 7595 7595 143916
F 3503 3503 140000
M 4092 4092
LOW 201045 201045 120000
F 143916 143916
100000
M 57127 57127
XNA 2 2
80000
MEDIUM 74046 74046
0
F 40859 40859 60000
57127
M 33185 33185
40859
XNA 2 2 40000 33185
20000
3503 4092
2 2
0
F M F M XNA F M XNA
HIGH LOW MEDIUM
From the above Bar plot we can infer that Females belonging to Low income group are the
highest number of clients with no payment issues
Target 1: Total_income_range vs Code_gender

Grand Total_income_range vs Code_Gender
Row Labels 1 Total 12000 11295
HIGH 468 468
F 180 180
10000
M 288 288
LOW 18551 18551
F 11295 11295 8000 7256
M 7256 7256
MEDIUM 5806 5806 6000
F 2695 2695 1
M 3111 3111
4000
Grand Total 24825 24825 3111
2695
2000
180 288
0
F M F M F M
HIGH LOW MEDIUM
From the above Bar plot we can infer that Females belonging to Low income group are the
highest number of clients with payment issues
Target 0: Credit Amt vs Education status

TARGET 0
Count of NAME_EDUCATION_TYPE Column Labels

Secondary /
Incomplete secondary
Row Labels Academic degree Higher education higher Lower secondary special Grand Total
HIGH 73 28670 2774 784 60996 93297
LOW 41 20434 3628 1516 73101 98720
MEDIUM 47 21750 3003 1099 64770 90669
Grand Total 161 70854 9405 3399 198867 282686
AMT_CREDIT vs EDUCATION STATUS From the adjacent Bar

Plot we can infer that
73101
clients having credit
64770
amt range as ‘Low’
60996
Academic degree
and education status
Higher education as ‘Secondary/
Incomplete higher Secondary Special’
Lower secondary have the highest count
28670 Secondary / secondary special for clients with no
20434 21750
payment issues
73 2774 784 41 3628 1516 47 3003 1099
HIGH LOW MEDIUM

Target 1: Credit Amt vs Education status

TARGET 1
Count of
NAME_EDUCATION_TYPE Column Labels
Secondary
/
Academic Higher Incomplete Lower secondary Grand
Row Labels degree education higher secondary special Total
HIGH 2 1397 205 74 4922 6600
LOW 1081 322 161 6878 8442
MEDIUM 1 1531 345 182 7724 9783
Grand Total 3 4009 872 417 19524 24825
AMT_CREDIT vs EDUCATION_STATUS From the adjacent Bar

Plot we can infer that
Academic degree clients having credit
7724
6878 Higher education
amt range as ‘Medium’
and education status
4922 Incomplete higher
as ‘Secondary/
Lower secondary Secondary Special’
have the highest count
Secondary / secondary
2 1397 205 74 1081 322 161 1 1531 345 182 for clients with
special
HIGH LOW MEDIUM payment issues
Target 0: Total Income vs Family status

TARGET 0
Count of NAME_FAMILY_STATUS Column Labels

Row Labels Civil marriage Married Separated Single / not married Unknown Widow Grand Total
HIGH 617 5248 507 1042 1 180 7595
LOW 19204 127305 12688 29530 12318 201045
MEDIUM 6993 49029 4955 10415 1 2653 74046
Grand Total 26814 181582 18150 40987 2 15151 282686
TOTAL_INCOME_RANGE vs FAMILY STATUS

140000
127305
120000
From the adjacent Bar plot
100000 we can infer that clients with
Civil marriage
total_income_range as ‘Low’
Married
80000
Separated
and family_status as
60000 Single / not married ‘Married’ have the highest
49029
Unknown count for clients having no
Widow
40000
29530 payment issues
19204
20000 12688 12318 10415
5248 6993 4955
617 507 1042 1 180 1 2653
0
HIGH LOW MEDIUM
Target 1: Total Income vs Family status

TARGET 1
Count of NAME_FAMILY_STATUS Column Labels

Row Labels Civil marriage Married Separated Single / not married Widow Grand Total
HIGH 46 301 46 70 5 468
LOW 2227 10998 1170 3400 756 18551
MEDIUM 688 3551 404 987 176 5806
Grand Total 2961 14850 1620 4457 937 24825
TOTAL_INCOME_RANGE VS FAMILY_STATUS
12000
10998
10000
From the adjacent Bar plot
8000 we can infer that clients with
Civil marriage
Married
total_income_range as ‘Low’
6000
Separated and family_status as
3551
Single / not married ‘Married’ have the highest
4000 3400 Widow
count for clients having
2227
2000
1170
payment issues
756 987
688
301 404 176
46 46 70 5
0
HIGH LOW MEDIUM
Google Drive Link for Excel sheet of Analysis of Cleaned Data

done:-
application_data_cleaned.xlsx - Google Drive

Previous Application Dataset – Dropping,
Imputing and analyzing Null values
The following columns of the previous application datasets need to be dropped as

they are irrelevant for doing the data analysis
• HOUR_APPR_PROCESS_START
• WEEKDAY_APPR_PROCESS_START_PREV
• FLAG_LAST_APPL_PER_CONTRACT
• NFLAG_LAST_APPL_IN_DAY
• SK_ID_CURR
• WEEKDAY_APPR_PROCESS_START
Removing the rows with the values 'XNA' &'XAP' for the column:
NAME_TYPE_SUITE
AMT_ANNUITY
Replace Blanks with 21340
Median of AMT_ANNUITY
21340
Previous Application Dataset – Dropping,
Imputing and analyzing Null values
NAME_TYPE_SUITE
Row Labels Count of NAME_TYPE_SUITE
Children 343
Family 3146
Group of people 39
Other_A 98
Other_B 276
Spouse, partner 1194
Unaccompanied 21206
(blank)
Grand Total 26302
Total count of each NAME_TYPE_SUITE

25000
21206
20000
15000
Replace Blanks with Unaccompained
10000
5000 3146 Total

343 276 1194
39 98
0
Previous Application Dataset – Analysis of
Cleaned Data
Distribution of Name Contract Status
Count of NAME_CONTRACT_STATUS Column Labels
Row Labels Approved Canceled Refused Unused offer Grand Total
Building a house or an annex 434 60 1188 1682
Business development 78 12 164 254
Buying a garage 28 5 51 84
Buying a holiday home / land 91 13 230 334
Buying a home 130 23 393 546
Buying a new car 139 29 465 4 637
Buying a used car 552 57 1166 9 1784
Car repairs 223 14 256 493
Education 481 14 476 4 975
Everyday expenses 732 8 740 7 1487
Furniture 210 15 250 475
Gasification / water supply 75 3 125 203
Hobby 11 20 31
Journey 329 10 404 2 745
Medicine 676 25 696 5 1402
Money for a third person 10 6 16
Other 4106 186 5310 62 9664
Payments on other loans 189 45 973 3 1210
Purchase of electronic equipment 357 4 280 3 644
Refusal to name the goal 1 7 8
Repairs 5385 381 8973 28 14767
Urgent needs 2228 83 2998 5309
Wedding / gift / holiday 248 10 336 594
Grand Total 16713 997 25507 127 43344
Previous Application Dataset – Analysis of
Cleaned Data
Distribution of Name Contract Status
Wedding / gift /…248
10
336
Urgent needs 2228 83 2998
Repairs 5385 381 8973 28
Refusal to name the…1
7
Purchase of…35742803
Payments on other…189
45 973 3
Other 4106 186 5310 62
Money for a third…
10
6
Medicine 676 25 696 5
Journey 329
104042
Hobby 11
20 Approved
Gasification / water…75
125
3 Canceled
Furniture 210
15
250
Refused
Everyday expenses 732 8 740 7
Unused offer
Education 48114476 4
Car repairs 223
14
256
Buying a used car 552 57 1166 9
Buying a new car 139
29465 4
Buying a home 130
23393
Buying a holiday…91
13
230
Buying a garage 28
51
5
Business development 78
12
164
Building a house or… 43460 1188
0 2000 4000 6000 8000 10000 12000 14000 16000
From the above Bar Plot we can infer that Name of Contract status i.e. Repairs work has the highest count of Approved Loans
Previous Application Dataset – Analysis of Cleaned
Data
Google Drive Link for Excel sheet of Analysis of Cleaned Data

done:-
previous data cleaned.xlsx - Google Sheets

Hence the analysis are being done on both datasets Applications Dataset and
Precious Applications Dataset
The following conclusions were drawn from the analysis done
• The proportion/percentage of the defaulters(target = 1) is around 8% and that of
non-defaulters(target = 0) is around 92%
• The Bank generally lends more loan to Female clients as compared to Males
clients as the count of Female clients in the defaulter’s list is less than that of
Males. Still Bank can look for more Male clients if their credit amount is satisfied
• Also the clients who belong to Working class tend to pay their loans on time
followed by the clients who fall under Commercial Associate
• Clients having Education status like Secondary/ Higher Secondary or more tend
to pay loan on time so bank can prefer lending loans to clients having such
Education Status
• Clients who fall in the Age Group 31-40 have the highest count for paying off their
loans on time followed by the clients who fall in the Age Groups 41-60
• Clients having LOW credit amount range tend to pay off their loans on time than
compared to HIGH and MEDIUM credit range
• Clients living with their Parents tend to pay off their loans quickly as
compared to other housing type. So Bank can lend loan to clients having
housing type → Living with Parents
• Clients taking loan for purchasing New Home i.e. clients taking Home Loans
or purchasing New Car i.e. Car Loans and clients who have a income type
as State Servant tend to pay their loans on time and hence Bank should
prefer clients having such background
• The Bank should be more cautious when lending money to clients with
Repairs purpose because they have high count of Defaulters along with
High count of Defaulters
Google Drive Folder Link for the Analysed datasets in form of Excel sheets
Due to vastness of data the Excel sheets needs to be downloaded and viewed
offline:-
trainity_task_6_final_project_2 - Google Drive

Trainity Data Analytics Trainee Task 6

Uploaded by

Copyright:

Available Formats

Trainity Data Analytics Trainee Task 6

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Trainity Data Analytics Trainee Task 6

Uploaded by

Copyright:

Available Formats

Name : ADVAIT GURUNATH CHAVAN

Analysis done on the following points:

To identify patterns which indicate if a client has difficulty paying their

Replacing Blanks in Occupation_Type column of the Application

Total count of each Occupation_type Accountants

Private service staff 2652

Replacing Blanks in AMT_ANNUTIY column of the Application Dataset

Replacing Blanks with Median

Replacing Blanks in AMT_GOODS_PRICE column of the Application

Replacing Blanks with Median

Replacing Blanks in Name_Type_Suite column of the Application

Replacing Blanks in Organization_type column of the Application

Dataset with the highest occurring categorical variable Cleaning

Total count of each Organization_type Highest occurring Emergency

‘Business Entity Type 3’ Industry: type 10

10553 10404 11193 Self-employed 38412

Trade: type 3 3492

First outlier is in AMT_ANNUITY

application_data_cleaned.xlsx - Google Drive

From the GENDER_VARIABLE pie chart

Row Labels Count of NAME_HOUSING_TYPE

Total count of each Housing Type

1122 11183 2617 4881 14840

own apartment i.e. the bank may consider the

living in Co-op apartment, Municipal Apartment,

Apartment and people living with their parents

0.36% 3.64% 0.85% 1.59% 4.83%

20-30 31-40 41-50 51-60 61-70

Clients Age Group with payment issues

20-30 31-40 41-50 51-60 61-70

Client amount credit range

Client amount credit range without payment

HIGH LOW MEDIUM

Client amount credit range

Client amout credit range with payment issue

From the adjacent Bar plot we can

HIGH LOW MEDIUM

BUSINESSMAN COMMERCIAL MATERNITY PENSIONER STATE SERVANT STUDENT UNEMPLOYED WORKING

COMMERCIAL MATERNITY LEAVE PENSIONER STATE SERVANT UNEMPLOYED WORKING

HIGH LOW MEDIUM

HIGH LOW MEDIUM

Univariate Analysis for TARGET variable

Count of CODE_GENDER Column Labels

Univariate Analysis for TARGET variable

Count of NAME_INCOME_TYPE Column Labels

10 5360 3 2 2982 20454

From the adjacent Bar Plot we can infer

Univariate Analysis for TARGET variable

ACADEMIC HIGHER INCOMPLETE LOWER SECONDARY /

Univariate Analysis for TARGET variable

defaulters i.e. CIVIL MARRIED SEPARATED SINGLE / NOT UNKNOWN WIDOW

Univariate Analysis for TARGET variable

Univariate Analysis for TARGET variable

From the adjacent Bar plot we can infer that clients

Bivariate Analysis for TARGET variable

Target 0: Total_income_range vs Code_gender

Bivariate Analysis for TARGET variable

Target 1: Total_income_range vs Code_gender

Bivariate Analysis for TARGET variable

Target 0: Credit Amt vs Education status