Advanced15 columnsUp to 10,000 rows Fictional data

Loan Applications dataset

Loan applications with income, credit score, loan size and term, and whether each was approved.

About this dataset

Applications to a fictional lender: who applied (age, education, employment), what they asked for (purpose, amount, term), their credit score and existing loans, and whether the loan was approved.

Approval was decided by a hidden scoring rule plus some randomness, the way real decisions mix rules with judgement. Your job is to discover the rule from the data. It rewards feature engineering: the raw columns help, but the EMI-to-income ratio you can work out yourself helps a lot more.

Good for practising: Classification, Feature engineering, Logistic regression, Model evaluation.

What's built into the data

The rows are random, but they follow rules, so there is something real to find. These are the patterns to look for:

Credit score and the EMI-to-income ratio are the two biggest drivers: a higher score helps, a heavier EMI hurts.
The EMI is not in the file. Work it out at 11% a year over the loan term and compare it with income.
Each existing loan lowers the odds; salaried applicants do a little better than others.
Loan sizes and terms depend on the purpose: home loans are the largest and longest, personal loans the smallest.

Columns

Column	Type	Description
`application_id`	ID (text)	Application number.
`applied_on`	Date	Date of application.
`age`	Whole number	Applicant's age.
`gender`	Category	Male or Female.
`marital_status`	Category	Married, Single or Divorced.
`education`	Category	Highest qualification.
`employment_type`	Category	Salaried, Self-employed, Business owner or Contract.
`annual_income`	Amount (₹)	Yearly income in rupees.
`loan_purpose`	Category	Home, Car, Personal, Education or Business.
`loan_amount`	Amount (₹)	Amount requested in rupees.
`loan_term_months`	Whole number	Repayment period in months.
`credit_score`	Whole number	Credit score, 300 to 900.
`existing_loans`	Whole number	Number of loans already running.
`city_tier`	Whole number	1 = metro, 2 = large city, 3 = smaller town.
`approved`	Category	Yes or No.

Preview: the first 8 rows

A sample with seed 2026. Your download will have different values but the same columns and patterns.

application_id	applied_on	age	gender	marital_status	education	employment_type	annual_income	loan_purpose	loan_amount	loan_term_months	credit_score	existing_loans	city_tier	approved
APP000001	2024-06-14	36	Female	Divorced	Graduate	Self-employed	545000	Personal	360000	12	616	0	1	No
APP000002	2024-06-26	35	Male	Married	Postgraduate	Salaried	428000	Personal	190000	36	681	0	3	Yes
APP000003	2024-09-10	21	Female	Married	Postgraduate	Salaried	1076000	Car	1130000	48	707	1	2	No
APP000004	2025-04-02	34	Female	Married	Graduate	Business owner	881000	Personal	350000	36	637	2	2	No
APP000005	2025-04-09	39	Female	Single	Graduate	Business owner	658000	Car	330000	60	631	3	1	No
APP000006	2025-05-12	21	Male	Married	Higher Secondary	Salaried	491000	Personal	480000	24	702	2	1	No
APP000007	2025-06-13	35	Male	Married	Graduate	Salaried	892000	Personal	390000	24	637	1	1	No
APP000008	2025-12-19	26	Female	Married	Graduate	Self-employed	681000	Car	440000	24	688	2	1	No

Practice exercises

Compute each applicant's EMI (11% a year, the given term) and EMI ÷ monthly income. Plot approval rate against it.
Train a logistic regression with and without your EMI-ratio feature. How much does it improve the model?
Plot approval rate by credit-score band (300–600, 600–700, 700–750, 750+).
Train a decision tree and read its first few splits. Do they match the patterns listed above?
Use a train/test split and report a confusion matrix and ROC-AUC, not only accuracy.

Stuck on one? Each exercise uses ideas from the lessons below, and trying it with the messy option on turns it into a cleaning exercise as well.

Load it straight into Python

You don't have to download first. Pandas can read the link directly, in Jupyter, Google Colab or a script. Keep the seed so you get the same data every time you run the notebook:

import pandas as pd

url = "https://beeprotonai.info/resources/datasets/loan-applications/download/?rows=1000&seed=42"
df = pd.read_csv(url, parse_dates=["applied_on"])
df.head()

For SQL practice, download the SQL format and run it against an empty database, for example sqlite3 practice.db < loan-applications-1000-rows-seed42.sql or mysql -u root -p practice < loan-applications-1000-rows-seed42.sql (use your file's name). It creates a table called loan_applications and fills it.

About the data

Everything in this file is invented. Names are random pairings of common first names and surnames, email addresses use example.com (reserved for examples, so it can never reach anyone), and phone numbers start with +91 00, which no real Indian mobile number does. Any resemblance to a real person or business is coincidence. Dates run from 1 April 2024 to 31 March 2026 unless a column says otherwise. Use it freely for learning, teaching and projects, including commercial ones, with no attribution required (CC0).

Lessons that go with it

More datasets

Sweet Shop Sales

Two years of orders from a fictional chain of Indian sweet shops, with Diwali …

Employee & Attrition Records

An HR file of employees with salary, satisfaction, overtime and whether they left: built …

Student Exam Results

Marks in five subjects with attendance and study hours, plus totals, percentage, result and …

Customer List (CRM)

Customer records with contact details, segment, orders and spend: ideal for cleaning, deduplication and …

Stock Register (Inventory)

Monthly stock register for five warehouses: opening, received, issued, closing, reorder level and value.

Website Traffic (Analytics)

Daily sessions, users, bounce rate and conversions by page, traffic source and device.

See all 14 datasets →