Dark Mode
Financial Credit Score Dataset
Finance & Banking Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset is designed for credit risk classification within the banking sector. It comprises customer transaction and demographic data, specifically categorising customers as either risky or not risky for particular banking products. The dataset aims to help new data scientists and data analysts explore Machine Learning and Statistical modelling concepts, addressing key questions such as identifying factors that contribute to a customer being credit risky and understanding the behaviour of credit worthy customers.
Columns
The dataset consists of two main files:
payment_data.csv for payment history and customer_data.csv for demographic data.payment_data.csv:
- id: A unique identifier for each customer.
- OVD_t1: The number of times a customer was overdue (type 1).
- OVD_t2: The number of times a customer was overdue (type 2).
- OVD_t3: The number of times a customer was overdue (type 3).
- OVD_sum: The total number of overdue days.
- pay_normal: The count of times a customer made a normal payment.
- prod_code: The specific credit product code.
- prod_limit: The credit limit associated with the product.
- update_date: The date when the account was last updated.
- new_balance: The current balance of the product.
- highest_balance: The highest balance recorded in the customer's history.
- report_date: The date of the most recent payment.
customer_data.csv:
- label: The target column indicating credit risk, where 1 denotes high credit risk and 0 denotes low credit risk.
- id: A unique identifier for each customer.
- fea_1 to fea_11: These are various demographic and behavioural features.
fea_1,fea_3,fea_5,fea_6,fea_7, andfea_9are encoded category features.
Distribution
The dataset is small and structured across two CSV files. The
customer_data.csv file contains 1125 records. Most columns in customer_data.csv, including label, id, fea_1, fea_3, fea_4, fea_5, fea_6, fea_7, fea_8, fea_9, fea_10, and fea_11, have data for all 1125 records without any missing values. However, fea_2 has 976 valid records with 149 missing entries.The
label column shows a distribution where 900 customers are in the low credit risk category (0) and 225 customers are in the high credit risk category (1).Usage
This dataset is ideal for:
- Developing and testing Machine Learning models for credit risk assessment.
- Statistical modelling of customer financial behaviour.
- Identifying key factors that contribute to a customer being deemed credit risky.
- Analysing patterns and characteristics of credit worthy customers.
- Building predictive models to classify whether a customer is Credit Risky or Credit Worthy from a banking perspective.
Coverage
The dataset primarily focuses on customer transaction and demographic data related to banking products. Specific details regarding the geographic origin, exact time range of the data, or a detailed breakdown of demographic groups are not provided. The dataset includes customer card payment history, account update dates, and recent payment dates, but a precise temporal scope is not specified.
License
CC BY-SA 4.0
Who Can Use It
This dataset is particularly useful for:
- Budding data scientists looking to gain practical experience in classification problems.
- Data analysts interested in exploring customer behaviour and financial risk.
- Researchers and developers working on E-Commerce Services and Lending applications.
- Anyone aiming to apply Machine Learning and Statistical modelling to real-world financial data challenges.
Dataset Name Suggestions
- Banking Customer Credit Risk
- Financial Credit Score Dataset
- Customer Payment Behaviour Analytics
- Credit Risk Prediction for Banking
- Banking Product Credit Assessment
Attributes
Original Data Source: Financial Credit Score Dataset
QUALITY