What's inside the box? An introduction to explainability in finance

Presentation: Explainability in finance

Workshop description

Building a machine learning model is just one piece of the puzzle in data science; explaining how it works is just as important, especially in finance where transparency and explainability are key.

During this workshop, we'll show you how to look inside your model and understand what’s going on. We'll explore techniques to make your models easier to explain and discuss their opportunities and downfalls.

File Structure

.   
├── solutions/ 					# Solutions directory   
│   ├── Explainability_solutions.ipynb		# Notebook with solutions to workshop  
│   └── Explainability_solutions_extended.ipynb	# Notebook with additional exercises and their solutions to explore by yourself
│  
├── workshop/					# Workshop directory  
│   ├── Explainability.ipynb			# Notebook with exercises  
│   ├── mortgage_churn.csv			# Data for workshop  
│   └── PyLadies_Explainable_AI.pdf		# Presentation
│   
├── README.md 									
└── requirements.txt				# List of dependencies  

Requirements

• Python 3.10 or higher
• Jupyter notebook

Usage

Jupyter option

• Clone the repository with git clone https://github.com/pyladiesams/intro-to-explainabilty-in-finance-oct2024
• Setting up your environment Install environment: python -m venv explainability_venv
  Activate environment: explainability_venv\Scripts\activate or source explainability_venv/bin/activate depending on OS
  Install dependencies: pip install -r requirements.txt
  Install notebook kernel for the venv: python -m ipykernel install --user --name explainability_venv
  Start notebook with jupyter notebook and select kernel with environment

Google colab option

• Open Collab notebook
• Create a copy of this notebook in your Google Drive. Go to Menu options: File > Save a copy in Drive. This will open a copy of the notebook in a new window.

Video record

Re-watch this YouTube stream

Credits

This workshop was set up by @pyladiesams, Magdalena Zych, Petra Gibcus, Amirsalar Molaei, Dana Tokmurzina, Ellie Nasiri, Dana van der Wende, Nihed Harrak.