Scribd Logo
Upload

Welcome to Scribd!

  • 15 views

    EDA Assignment Summary PDF

    Uploaded by

    Shikha
    The document discusses analyzing loan application data to determine factors that lead to default. It describes cleaning the data, then performing univariate and bivariate analysis on variables like loan grade, term, purpose, amount and income to identify relationships with default rates and segment loan categories by risk.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    EDA Assignment Summary PDF

    Uploaded by

    Shikha
    15 views12 pages
    The document discusses analyzing loan application data to determine factors that lead to default. It describes cleaning the data, then performing univariate and bivariate analysis on variables like loan grade, term, purpose, amount and income to identify relationships with default rates and segment loan categories by risk.

    Original Title

    EDA Assignment summary.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses analyzing loan application data to determine factors that lead to default. It describes cleaning the data, then performing univariate and bivariate analysis on variables like loan grade, term, purpose, amount and income to identify relationships with default rates and segment loan categories by risk.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    15 views12 pages

    EDA Assignment Summary PDF

    Uploaded by

    Shikha
    The document discusses analyzing loan application data to determine factors that lead to default. It describes cleaning the data, then performing univariate and bivariate analysis on variables like loan grade, term, purpose, amount and income to identify relationships with default rates and segment loan categories by risk.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 12

    UpgradProject

    Statistics & EDA

    By :- Shikha
    The Given data hasinformationontheloanapplication.The data given below
    contains the information about pastloan applicants and whether they
    ‘defaulted’ or not.
    The main objective of the analysis was to determine the conditions and situations that
    leads to an applicant being charged off or default.
    For this task, we need to predict the “loan_status” column of the dataset which specifies
    the status of the loan .
    The datasethadinitially 111 columns with 39716 entries.
    Consumer finance company-largest online loan marketplace, facilitating personal loans,
    business loans, and financing of medical procedures.
    2 risks associated with banks decision to approve loans.
    1. loss of likely to repay the loan, then not approving the loan results in a loss of business to
    the company
    2. not likely to repay the loan, then approving the loan may lead to a
    financial loss for the company
    This task will be done by using univariate and bivariate analysis of different
    columns of the dataset.

    So, naturally the first step would be to reduce these to a sizable quantity.
    Now, of the 28 columns we need to find the ones which affect
    the target variable ‘loan_status’. We’ll do this by comparing it
    with other columns and by analyzingeach of these columns
    on their own.
    On analyzingour ‘loan_status’ column we find that the overall
    default rate is up to 14%
    To start things off, let’s look at all the categorical columns
    first.
    We will plot them against our target variable ‘loan_status’.
    It can be clearly seen that the risk of loan increases as we go
    from grade A to G, which is expected because of LC
    guidelines of assigning the grade.
    From this it can be observed
    that loans of 60 months term
    tend to default more than 36
    months term loans.
    Plotting the purpose of loans
    shows that small business, debt
    consolidation, educational and
    renewable energy loans default
    more than any other category.
    After analyzingcategorical
    variables let’s now move on to
    continuous variables. We will bin
    these variables into different
    categories to plot them better.

    Loan amount shows that as loan


    amount increases loan tend to
    default more.
    Annual income seems to
    inversely affect the default rate.
    Which is an expected result.
    Now let’s do some segmented univariate
    analysis.
    Let’s analyzevarious categories with
    ‘loan_status’, segmenting it with purpose
    First let’s compare it with term.
    Interesting,
    From above plots, a general
    trend can be observed that the
    debt consolidation loans have
    higher default rates in almost
    every other category.
    Segmented univariate analysis doesn’t yield much insight but it can be seen that debt
    consolidation loans tends to default
    frequently. The factors to consider while giving
    loans are: • Annual Income
    • Term
    • Interest Rate
    • Grade
    • Loan Amount
    • Purpose

    You might also like