Banking Portfolio Performance Dataset
A portfolio’s overall delinquency rate can improve even when the rate within each customer segment gets worse. This dataset lets you examine how that happens.
Download synthetic lending-account records and compare August and September 2026. Separate changes within customer segments from changes in the mix of accounts in the portfolio.
About the dataset
Created for credit-risk analysts, portfolio analysts and data scientists, this exercise includes 1,600 fictional accounts, 2,000 monthly account observations and two customer segments.
There are 1,000 observed accounts in each month. Some accounts remain, some leave and others enter. Your task is to explain the headline change using the underlying records.
Dataset details
| Industry | Banking |
| Formats | Excel workbook and CSV files |
| Structure | Three related data tables |
| Source records | 3,602 |
| Fields | 10 across the three table schemas |
| Observation dates | August 31 and September 30, 2026 |
| Currency | USD |
| Source | Original synthetic data created for Jolene.pro Workbook Studies |
| License | Creative Commons Attribution 4.0 International (CC BY 4.0) |
What’s included
Accounts: 1,600 records containing account IDs, fixed segment IDs and origination dates.
Snapshots: 2,000 records containing account IDs, observation dates, outstanding balances and days past due.
Segments: Two records defining the customer segments.
The complete ZIP includes the Excel workbook, three matching CSV files, a data dictionary, scenario notes, a preview and license details.
Recommended analysis
- Did the overall 30+ days-past-due account rate improve between August and September?
- Did the rate within each customer segment improve?
- What would September’s account rate be using August’s segment weights?
- How many accounts entered, exited or remained in the portfolio?
How to work with the data
Join Snapshots to Accounts using account_id, then join Accounts to Segments using segment_id.
For this exercise, an account is delinquent when days_past_due is 30 or more. Calculate account-based rates using account counts. For balance-based rates, use outstanding principal balances instead. Keep those denominators separate.
A fixed-weight comparison helps describe the effect of changing portfolio composition. It does not establish a causal explanation.
You can explore the workbook in Excel or Jolene.pro, or use the CSV files with SQL, Power BI and other analysis tools.
Downloads
Excel workbook and complete dataset ZIP.
Reuse and attribution
These are synthetic records, not customer data or industry benchmarks. You may reuse the dataset under CC BY 4.0. Credit Jolene.pro and identify any changes you make.