Silvercat Credit Risk Modelling Exercise

This repository contains a light-touch end-to-end credit risk modelling exercise using Lending Club loan application data.

The work is structured across three notebooks:

1. 1_analytics.ipynb
   Exploratory analysis, target definition, univariate risk analysis and portfolio composition review.

2. 2_modelling.ipynb
   Feature preparation, model development, tuning, validation and strategy simulation.

3. 3_deployment.ipynb
   Packaging of the fitted model and preprocessing logic into a reusable inference function.

The notebooks are logically sequenced and are intended to be read in order. However, they have been designed so that each notebook can be run independently from the repository, using the shared helper functions in src/silvercat/.

Repository contents

All supporting project assets are stored in this repository, including:

• notebooks
• helper Python modules
• modelling and inference code
• requirements and project configuration
• data dictionary / technical exercise files, where applicable

The only asset not stored in the repository is the Lending Club loans dataset, due to file size. The dataset is downloaded separately when running the notebooks.

Running the notebooks

Each notebook starts by setting up the project environment and loading the required helper functions. The expected workflow is:

from silvercat.utils import setup

paths = setup(project_root="/content/silvercat")