Credit Risk Analytics in 5 Days: Build a Scorecard in Python (Self-Study Bundle)

By the end of this course you will have built a complete application scorecard yourself: you will have defined good and bad loans, binned the variables, calculated Weight of Evidence and Information Value, fitted a logistic model, scaled it to points, measured it with Gini and KS, and checked it for drift with PSI. You will be able to explain each step in plain words.

This is a self-study bundle. You download one zip file and work through it at your own pace. The "5 days" are five lessons inside the course PDF. Do one a day, or all of them in a weekend.

WHAT'S INSIDE

• Course bundle PDF (37 pages): five lessons with worked examples on the synthetic data, a glossary, and notes on the data.
• Day 1: How a loan lives and dies. Lending lifecycle, DPD buckets, RBI SMA/NPA classes, expected loss (PD × LGD × EAD).
• Day 2: Define "bad" before you build. Observation and performance windows, indeterminates, out-of-time testing, leakage.
• Day 3: Binning, Weight of Evidence and Information Value in Python.
• Day 4: Fit the scorecard, scale it to points, score one applicant by hand, and measure it with Gini and KS.
• Day 5: Validate and monitor. PSI, CSI, cut-offs, calibration, reject inference, and how to explain the project in an interview.
• Printable workbook PDF (30 A4 pages): key ideas, exercises and a self-check for each day, plus a glossary.
• Answer key PDF: answers to every exercise and self-check.
• Synthetic dataset (CSV, CC0): 5,000 personal loans with 12-month outcomes, plus 2,000 recent applications for drift checks. It comes with a full data dictionary. Every row is computer-generated.
• Jupyter notebook and Python scripts: these reproduce every number in the lessons. Use them to check your work or as a starting point for your own version.
• START-HERE guide: what each file is, and how to set up Python or use Excel instead.

WHO IT'S FOR

• Students and freshers who want to understand how credit scorecards are actually built.
• Analysts moving into credit risk, lending or fintech analytics.
• Anyone preparing to talk through a scorecard project, who wants to have built one end to end.

WHO IT'S NOT FOR

• Experienced model validators and credit risk modellers. This covers the fundamentals.
• Anyone looking for a ready-made model to use on real lending decisions.
• Anyone looking for live classes, mentoring or one-to-one help. This is a self-study download.

REQUIREMENTS

• Basic Python (pandas) or basic Excel / Google Sheets. Days 1 to 3 work in a spreadsheet. Days 4 and 5 are easier in Python.
• Python 3.10 or newer if you want to run the notebook. Google Colab works.
• A PDF reader. A printer is optional, for the workbook.

PLEASE NOTE

• All data is synthetic. No real borrower, lender, bureau or employer data is used.
• Educational only. This is not financial, legal, credit or regulatory advice.
• No guarantees. The course does not promise a job, an interview, a salary or any other outcome.
• Delivery: an instant download of one zip file after purchase. No email course or drip is included. All five lessons are in the PDF. All sales are final, as this is an instant digital download. If the file doesn't arrive or won't open, contact us and we will fix it.
• Licence: for your personal learning use. Please don't resell or share the files. The datasets are CC0. The fonts are SIL Open Font License.