National Student Loan Data System

The dataset, sourced from the National Student Loan Data System, provides a comprehensive overview of student loan information for various educational institutions across the United States.

https://media.giphy.com/media/mFlGEZllc8QBJJAms2/giphy.gif" alt="img">

The data covers loan recipients, origination and disbursement counts, as well as the corresponding monetary values. It spans the academic years from 2009 to 2010, including all four quarters of each year.

https://media.giphy.com/media/og9wDNldG2M4rd13pN/giphy.gif" alt="img">

The dataset is invaluable for understanding loan patterns, disbursement trends, and recipient demographics within different educational institutions during this period.

https://media.giphy.com/media/iH1rdIKYvNhOu5Z8RY/giphy.gif" alt="">

The columns are :

OPE ID: - Definition: An 8-digit code uniquely identifying the school at its main branch. - Significance: Provides a specific identifier for each educational institution for accurate tracking and database management.

School: - Definition: The name of the educational institution associated with the OPE ID. - Significance: Identifies the specific school or university corresponding to the loan data, allowing for institution-specific analysis.

State: - Definition: The state where the main campus of the educational institution is located. - Significance: Offers geographical context, enabling regional comparisons and understanding loan dynamics in different states.

Zip Code: - Definition: The zip code of the main campus of the educational institution. - Significance: Provides specific location data, enhancing the granularity of the dataset and allowing for localized analysis.

School Type: - Definition: Indicates the control or ownership of the school (e.g., PRIVATE, PUBLIC). - Significance: Classifies schools based on ownership, enabling distinctions in loan trends between private and public institutions.

Recipients (Q1, Q2, Q3, Q4): - Definition: The number of loan recipients for the specified loan type during each quarter of the academic year. - Significance: Reflects the count of students or individuals receiving loans, providing a quarterly breakdown of loan recipients.

# of Loans Originated (Q1, Q2, Q3, Q4): - Definition: The number of loans initiated for the specified loan type during each quarter of the academic year. - Significance: Indicates the count of new loans originated during each quarter, offering insights into borrowing trends over time.

$ of Loans Originated (Q1, Q2, Q3, Q4): - Definition: The total dollar amount of loans initiated for the specified loan type during each quarter of the academic year. - Significance: Highlights the financial magnitude of loan originations for each quarter, showcasing the monetary aspects of borrowing.

# of Disbursements (Q1, Q2, Q3, Q4): - Definition: The number of disbursements made for the specified loan type during each quarter of the academic year. - Significance: Indicates the frequency of fund allocations, portraying the administrative workload related to disbursements each quarter.

$ of Disbursements (Q1, Q2, Q3, Q4): - Definition: The total dollar amount of disbursements made for the specified loan type during each quarter of the academic year. - Significance: Represents the cumulative disbursement amount for each quarter, providing insights into the financial distribution of student loans over time.

Dataset

This dataset was created by Sifat Ibti

Released under GNU Free Documentation License 1.3

Contents

All files needed to replicate data analysis and figures in "The Political Benefits of Student Loan Debt Relief"

Over 44.7 million Americans carry student loan debt, with the total amount valued at approximately $1.31 trillion (Quarterly Report, 2019). Ergo, consumer spending, a factor of GDP, is stifled and negatively impacts the economy (Frizell, 2014, p. 22). This study examined the relationship between student loan debt and the probability of a recession in the near future, as well as the effects of proposed student loan forgiveness policies through the use of a created model. The Federal Reserve Bank of St. Louis’s website (FRED) was used to extract data regarding total GDP per quarter and student loan debt per quarter ("Federal Reserve Economic Data," 2019). Through the combination of the student loan debt per quarter and total GDP per quarter datasets, the percentage of total GDP composed of student loan debt per quarter was calculated and fitted to a logistic curve. Future quarterly values for total GDP and the percentage of total GDP composed by student loan debt per quarter were found through Long Short Term Models and Euler’s Method, respectively. Through the creation of a probability of recession index, the probability of recession per quarter was compared to the percentage of total GDP composed by student loan debt per quarter to construct an exponential regression model. Utilizing a primarily quantitative method of analysis, the percentage of total GDP composed by student loan debt per quarter was found to be strongly associated[p < 1.26696* 10-8]with the probability of recession per quarter(p(R)), with the p(R) tending to peak as the percentage of total GDP composed of student loan debt per quarter strayed away from the carrying capacity of the logistic curve. Inputting the student loan debt forgiveness policies of potential congressional bills proposed by lawmakers found that eliminating 49.7 % and 36.7% of student loan debt would reduce the recession probabilities to be 1.73545*10-29% and 9.74474*10-25%, respectively.

https://raw.githubusercontent.com/Masterx-AI/Project_Loan_Default_Risk_Expectancy_/main/loan.jpg" alt="">

Description:

Banks earn a major revenue from lending loans. But it is often associated with risk. The borrower's may default on the loan. To mitigate this issue, the banks have decided to use Machine Learning to overcome this issue. They have collected past data on the loan borrowers & would like you to develop a strong ML Model to classify if any new borrower is likely to default or not.

The dataset is enormous & consists of multiple deteministic factors like borrowe's income, gender, loan pupose etc. The dataset is subject to strong multicollinearity & empty values. Can you overcome these factors & build a strong classifier to predict defaulters?

Acknowledgements:

This dataset has been referred from Kaggle.

Objective:

• Understand the Dataset & cleanup (if required).
• Build classification model to predict weather the loan borrower will default or not.
• Also fine-tune the hyperparameters & compare the evaluation metrics of vaious classification algorithms.

The National Postsecondary Student Aid Survey (NPSAS) provides information on how postsecondary student financial aid is targeted, received, and used. A significant component of the NPSAS is the Student Loan Recipient Transcript Survey, which collected postsecondary-school transcripts for Guaranteed Student Loan (GSL) recipients who were surveyed in the Student Loan Recipient Survey (SLRS, Part 4). This component provides the means to analyze basic policy issues such as relationships between educational activities and ability to cope with indebtedness, and the patterns of student loan repayment or default. The Transcript Survey data cover 11,847 students, 12,213 transcripts, and 1,412 schools and are organized into four categories, consisting of data at the student (Part 5), transcript (Part 7), term (Part 6), and course (Part 2) levels. At least one student-level and one transcript-level record exist for each sample member for whom a transcript was requested, even if the school in question reported that an individual had never attended or had withdrawn before establishing a formal record of attendance. Student-level data (Part 5) provide general information about the respondent's academic career. Each record is given a case ID code, allowing the merger of transcript data and other files, sampling weights, and data that summarize information found on transcripts from all postsecondary schools attended as well as selected items from Part 4, the Student Loan Recipient Survey data files. Transcript-level records (Part 7) contain data pertaining to the student's academic record at a single institution, including the school ID code, degree or other credentials conferred with accompanying dates, major and minor field(s) of study, and the student's cumulative grade-point average. Term records (Part 6) contain type of term (quarters, trimesters, and semesters), season of occurrence, start and end dates, grade-scale type, the number of courses corresponding to a given term, and a special flag indicating regular or transfer status for the term. Included in term type is a code that signifies credit earned via standardized tests and other life experience. Course-level data (Part 2) include records for every course reported on a transcript. The Student Loan Recipient Survey data (Part 4, Questionnaire Data) contain identifying information about the students such as sex, age, race, citizenship, residence, marital status, and current employment, as well as survey control data, a counter variable for the NPSAS transcripts, and weights. The Composite Data file (Part 1) contains information from the student transcript data in Part 5 and the student questionnaire data in Part 4. It also contains composite variables that combine information from the record abstract done at the institution and the student questionnaire. A Parent Survey (Part 3) was also conducted to collect data on the total number of children in the family, how much respondents spent on clothing, food, and books and supplies for their children, other loans taken out to pay for schooling, when the respondents started saving for their children's college expenses, and what type of savings programs they used.

Required qualitative dataset for Ph.D. thesis submission: · The dataset consists of forty-one respondents who participated in in-depth semi-structured interviews. · Respondents were recruited from eight educational institutions: three 4-year public universities, two 3-year public technical colleges, two 4-year private colleges (independent second-tier schools), and one 3-year private vocational college. · In detail, twenty-two respondents were female, and four were from minority ethnic groups. · All interviews were audio-recorded and later transcribed. · Interviews were conducted in Chinese, and Word files ( .docx) transcripts were imported into NVivo 11 Plus (Windows) to create nodes (two-layer coding), memos, and thematic maps in NVivo format through reflexive thematic analysis.

This 1989 survey of students in schools and FE colleges was conducted by the National Union of Students to measure the likely effects on students' intentions regarding Higher Education of the introduction of student loans. One of the NUS researchers had previously attended the Survey Analysis Workshop course (post-grad, part-time, evening) offered by the Survey Research Unit (SRU) Polytechnic of North London. She approached John Hall (Course Tutor and SRU Director) about the feasibility of a survey to provide evidence for the NUS campaign against student loans. The NUS constructed a random sample of schools and FE colleges from their register: questionnaires were distributed to those agreeing to take part and were administered by the institutions themselves, some before the announcement, others afterwards. Advice on questionnaire design, and technical assistance with data processing and analysis were provided by John Hall. Data processing and analysis of the survey was done at the Polytechnic of North London. See: Aughterson K and Foley K 'Opportunity Lost: a survey of the intentions and attitudes of young people as affected by the proposed system of student loans' National Union of Students, 1989.

Dataset Description

The dataset includes 1,000 records with information about loan applications, including variables related to the applicant's financial status, credit history, and loan details. The goal is to analyze patterns in credit risk or build models to predict loan defaults.

Columns:

• checking_balance: Customer's current account balance in deutschmarks, classified as < 0 DM (negative balance), 1 - 200 DM, > 200 DM, or unknown (unknown).
• months_loan_duration: Duration of the loan in months.
• credit_history: Credit history of the applicant.
• purpose: Purpose of the loan.
• amount: Loan amount.
• savings_balance: Savings account balance.
• employment_duration: Length of employment.
• percent_of_income: Percentage of income allocated to loan repayment.
• years_at_residence: Years at the current residence.
• age: Applicant's age.
• other_credit: Presence of other credit agreements.
• housing: Housing status (e.g., rent, own).
• existing_loans_count: Number of existing loans.
• job: Job type or classification.
• dependents: Number of dependents.
• phone: Availability of a telephone.
• default: Target variable indicating loan default ("yes" or "no").

Inspiration

This dataset can be used for: - Building predictive models for loan default. - Exploring relationships between financial variables and credit risk. - Enhancing your understanding of credit risk analysis.

License

This dataset is published under the CC BY-NC-SA 4.0 license: - Permitted: Educational, research, and personal use. - Restricted: Commercial use is not allowed. - Attribution: Credit to Universidad de Santiago de Chile is required. - Sharing: Derivative works must use the same license.

This dataset was originally provided by the Universidad de Santiago de Chile as part of the course "Machine Learning for Management". I am not the original creator of the data, and my role is solely to share this resource for educational and research purposes. All rights to the original data belong to the university and/or the original authors.

This dataset may not be used for commercial purposes or in contexts that violate the copyright or policies of the institution that created it. Users are responsible for complying with the terms of use specified in the accompanying license and should ensure they provide appropriate credit.

Additional Notes If you are a student or researcher interested in using this dataset, please make sure to give proper credit to the original source in your publications or projects.

The Survey of Consumer Finances (SCF) is normally a triennial cross-sectional survey of U.S. families. The survey data include information on families' balance sheets, pensions, income, and demographic characteristics.

The Student Income and Expenditure Survey (SIES) is designed to collect detailed information on income and expenditure of Higher Education students, and investigates issues such as student debt or hardship. The survey covers both full-time and part-time students at higher education institutions (HEI) and further education colleges (FEC), including the Open University (OU), participating in undergraduate courses. Undergraduate courses included first degree and Higher National Diplomas/Certificates (HNDs/HNCs), or in university-based postgraduate initial teacher training courses (PGCEs).

The 2011/12 survey is the latest in a series of surveys carried out at approximately three year intervals. The methods and interview content have been kept as similar as possible to previous waves in order to make any trend comparisons as robust as possible.

The main aims of the SIES 2011/12 Survey were to:

• provide detailed information on the income, expenditure and debt levels of higher education (HE) students in England and Wales
• allow for analysis on larger and more memorable spending captured in the main questionnaire, as well as day-to-day spending recorded in the seven-day spending diary
• provide a baseline for assessing the impact of changes in student finance introduced in September 2012 for those starting HE in the 2012/13 academic year

Fieldwork was conducted between February 2012 and June 2012. Please see the User Guide accompanying the SIES 2011/12 dataset for further information.

Secure Access Dataset and Related Studies:
In the Secure Access version of SIES 2011/12 the raw financial variables have not been banded, as was the case for the standard End User Licence (EUL) version held by the UK Data Archive under SN 7611. The Archive also holds an EUL version of SIES 2007/08 under SN 6319.

This report summarizes the findings of the Consortium's third annual survey, which involved 25 colleges and more than 9,400 students. Participating colleges were responsible for sampling (based on a standardized procedure) and administering the survey in class. Completed questionnaires were then shipped to PRA Inc. for coding, data entry and analysis. The objectives of the research are to: provide national data on student access, time use and financing for Canadian college students from participating colleges; identify issues particular to certain learner groups or regions; and provide each institution with topline survey results (based on representative samples of their students), which may then be compared against the "national average". This dataset was freely received from the Canada Millennium Scholarship Foundation. Some work was required for the variable and value labels, and missing values. They were corrected as best as possible with the documentation received. Caution should be used with this dataset as some variables are lacking information.

The Measuring the Effectiveness of Student Aid (MESA) dataset comprises a sample of low income students receiving student financial aid in 2006-07. Students were contacted first (Cycle I) in February-May of that academic year (the precise date varying by province), and were then followed up in 2007-08 (Cycle II), contacted in February April of that year. Students will be contacted again in 2008-09 for the last time. The dataset represents a national sample, including all provinces-except for Prince Edward Island. In the spring and summer of 2005, the Canada Millennium Scholarship Foundation negotiated a series of agreements with provincial governments to deliver a set of bursaries (known as “Access Bursaries”) to first-time, first-year undergraduates from low-income families. These agreements are all broadly similar though eligibility criteria vary slightly by jurisdiction (section 1, below, describes the Access Bursaries as they exist in each province). Students do not need to apply for the award separately; instead, they are automatically considered for the award through their application for provincial student assistance. The sample represents a particular subset of the students who received student financial aid in their first year of postsecondary education in 2006 07. In the majority of the provinces, this subset consists of the students who received a Low Income Bursary from the Millennium Scholarship Foundation. In British Columbia and Nova Scotia, a control group made up of students who received financial aid but not the Millennium Bursary was surveyed as well. The Ontario sample is made up those Millennium Bursary recipients who also received a Canada Access Grant and those who did not, with sub-samples selected from each group (all appear together in the data but can be separately identified). The Bursary and the Grant are awarded in similar amounts, but the eligibility requirements are different. This dataset was freely received from the Canada Millennium Scholarship Foundation. Some work was required for the variable and value labels, and missing values. They were corrected as best as possible with the documentation received. Caution should be used with this dataset as some variables are lacking information.

Dataset and replication package for the paper Temporal Discounting in Software Engineering: A Replication Study (Fagerholm, F., Becker, C., Chatzigeorgiou, A., Betz, S., Duboc, L., Penzenstadler, B., Mohanani, R., Venters, C. (2019). Temporal Discounting in Software Engineering: A Replication Study. 13th ACM/IEEE International Symposium of Empirical Software Engineering and Measurement (ESEM 2019)). The dataset consists of answers to a questionnaire on temporal discounting in a technical debt context. Two questionnaire templates illustrate how to gather the data for professional and student participants. An analysis script is provided which shows the details of the calculations and analyses performed for the paper. More information is given in the description file.

The Survey of Consumer Finances (SCF) dataset, provided by the Federal Reserve, offers comprehensive insights into the financial condition of U.S. households. This dataset is invaluable for researchers, policymakers, and analysts interested in understanding consumer behavior, wealth distribution, and economic trends in the United States.

The SCF dataset includes detailed information on household income, assets, liabilities, and various demographic characteristics. It is collected every three years and serves as a crucial resource for analyzing the financial well-being of American families.

Key Features: Income Data: Information on various sources of income, including wages, investments, and government assistance. Asset Ownership: Detailed accounts of household assets, such as real estate, retirement accounts, stocks, and other investments. Liabilities:Comprehensive details on household debts, including mortgages, credit card debts, and student loans. Demographics: Data covering age, education, race, and family structure, allowing for nuanced analysis of financial trends across different segments of the population.

Use Cases: Economic research and analysis, Policy formulation and assessment, Understanding wealth inequality, Consumer behavior studies

Citing the Dataset:

When using this dataset in your research, please ensure to cite the Federal Reserve Board and the SCF as the original source.

Note: The dataset is intended for educational and research purposes. Users are encouraged to adhere to ethical guidelines when analyzing and interpreting the data.

Context

The Consumer Financial Protection Bureau (CFPB) is a federal U.S. agency that acts as a mediator when disputes arise between financial institutions and consumers. Via a web form, consumers can send the agency a narrative of their dispute. An NLP model would make the classification of complaints and their routing to the appropriate teams more efficient than manually tagged complaints.

Content

A data file was downloaded directly from the CFPB website for training and testing the model. It included one year's worth of data (March 2020 to March 2021). Later in the project, I used an API to download up-to-the-minute data to verify the model's performance.

Each submission was tagged with one of nine financial product classes. Because of similarities between certain classes as well some class imbalances, I consolidated them into five classes:

• credit reporting
• debt collection
• mortgages and loans (includes car loans, payday loans, student loans, etc.)
• credit cards
• retail banking (includes checking/savings accounts, as well as money transfers, Venmo, etc.)

After data cleaning, the dataset consisted of around 162,400 consumer submissions containing narratives. The dataset was still imbalanced, with 56% in the credit reporting class, and the remainder roughly equally distributed (between 8% and 14%) among the remaining classes.

Acknowledgements

• Many thanks to Author halpert3 for providing us the Clean Dataset- Github URL.
• Data is organised by CFPB.

This dataset was generated from a public earning's call (press release article). And used to generate examples of the way real humans would speak regarding the matters in the article, within real world scenarios. Here they are below:

Here are the linguistic variations for each of the queries in the dataset, based on the example article provided:

Here are five examples related to strong average loan growth in US Personal Banking (#5):

1. Mortgage Loans: An increase in demand for mortgage loans contributed to the strong average loan growth in US Personal Banking. Customers taking advantage of low interest rates led to a surge in mortgage applications and approvals.

2. Auto Loans: Robust consumer spending and increased car sales led to higher demand for auto loans, contributing to the strong loan growth in US Personal Banking. Customers seeking financing options for purchasing vehicles played a significant role in this growth.

3. Personal Loans: The availability of personal loans with favorable terms and competitive interest rates attracted borrowers, resulting in strong average loan growth in US Personal Banking. Customers availed personal loans for various purposes such as home improvements, debt consolidation, or financing other personal expenses.

4. Small Business Loans: US Personal Banking also witnessed strong loan growth due to increased lending to small businesses. As entrepreneurs and small business owners sought capital for expansion, equipment purchases, or working capital, the demand for small business loans rose, contributing to the growth.

5. Student Loans: The higher education sector continued to rely on student loans to finance tuition fees and related expenses. With the increasing cost of education, a rise in student loan applications and approvals contributed to the strong average loan growth in US Personal Banking.

General Queries Query: "What was the revenue for Personal Banking and Wealth Management (PBWM) in the last quarter?"

Variation 1: "What were the PBWM revenues in the previous quarter?" Variation 2: "Can you provide the revenue figure for PBWM in the last quarter?" Variation 3: "How much revenue did PBWM generate in the last quarter?" Variation 4: "What was the total revenue for PBWM in the most recent quarter?" Variation 5: "Could you tell me the revenue earned by PBWM in the last quarter?" Query: "What were the revenue figures for different divisions under US Personal Banking?"

Variation 1: "Can you provide the revenue breakdown for various divisions within US Personal Banking?" Variation 2: "What were the revenues generated by the different divisions in US Personal Banking?" Variation 3: "How did the revenue distribution look across different divisions in US Personal Banking?" Variation 4: "What were the individual revenue figures for each division within US Personal Banking?" Variation 5: "Could you give me a breakdown of the revenues for different divisions in US Personal Banking?" Query: "How did operating expenses change for PBWM?"

Variation 1: "What was the change in operating expenses for PBWM?" Variation 2: "Were there any fluctuations in the operating expenses of PBWM?" Variation 3: "How did the operating expenses for PBWM evolve over the specified period?" Variation 4: "Can you provide insights into the changes in operating expenses for PBWM?" Variation 5: "What was the percentage change in operating expenses for PBWM?" Query: "What factors contributed to the increase in PBWM's cost of credit?"

Variation 1: "What were the drivers behind the rise in PBWM's cost of credit?" Variation 2: "Which factors influenced the increase in PBWM's cost of credit?" Variation 3: "Can you identify the elements that led to the higher cost of credit for PBWM?" Variation 4: "What were the contributing factors to the cost of credit escalation in PBWM?" Variation 5: "What were the key reasons behind the growth in PBWM's cost of credit?" Query: "What led to the decrease in PBWM's net income?"

Variation 1: "What were the factors responsible for the decline in PBWM's net income?" Variation 2: "Can you identify the causes of the reduction in PBWM's net income?" Variation 3: "What influenced the decrease in net income for PBWM?" Variation 4: "Were there specific drivers that contributed to the decline in PBWM's net income?" Variation 5: "What were the primary reasons behind the decrease in PBWM's net income?" These linguistic variations provide different ways to ask the same questions, allowing for a more diverse and robust training dataset for the chatbot.

Here are the extracted entities from the provided article:

Account Line Entities:

Revenues Operating expenses Cost of credit Net income Business Line Entities:

Personal Banking and Wealth Management (PBWM) Branded Cards Retail Services Retail Banking Global Wealth Management Markets Banking Investment Banking Corporate Lending...