labrijisaad / Technical-Test-AXA-Direct-Assurance

Data Science Technical Test Solution.

Data Science Technical Test Solution (AXA Direct Assurance)

⭐ This repo holds my solution for the Data Science Technical Test from AXA Direct Assurance's DATA/AI team.

Problem Objective

🚀The primary objective of this project is to train a model using historical data and then tackle the challenge of data drift in machine learning. Data drift refers to the degradation of model performance when applied to new data. The core goals include visualizing data drift and utilizing statistical methods like Jensen-Shannon divergence and Kolmogorov-Smirnov tests to accurately detect and quantify the extent of data drift.

Project Structure

The project is structured as follows:

• data/: This directory contains the necessary data files for the analysis. (it is excluded from version control using .gitignore.)
• notebooks/: This directory contains Jupyter notebooks with the code for the data analysis.
• results/: This directory contains the output visualizations generated during the analysis.
• README.md: This file provides an overview of the project and instructions on how to reproduce the analysis.
• .gitignore: This file contains the list of files that are ignored by git.
• requirements.txt: This file contains the list of packages that are required to run the analysis.
• .env: This file contains the environment variables used in the analysis.

Getting Started

To get up and running with the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/labrijisaad/data-science-technical-test.git

2. Navigate to the project directory:

   cd data-science-technical-test

3. Ensure Python 3.9 is installed. The required packages are listed in the requirements.txt file.

4. Create a virtual environment (highly recommended to maintain project isolation):

   python3 -m venv venv

5. Activate the virtual environment:

   • macOS/Linux:

   source venv/bin/activate

   • Windows:

   venv\Scripts\activate

6. Install dependencies listed in requirements.txt:

   pip install -r requirements.txt

7. Place the dataset files within the data/ directory.

8. Fire up the Jupyter Notebook server:

   jupyter notebook

9. Select the Relevant Notebook:

   Open the appropriate notebook from the notebooks/ directory:

   • labri_solution.ipynb: My proposed solution.
   • TestTechniqueDerive_Sujet.ipynb: The notebook with the original problem statement.