Lifetime Value Model

Estimating lifetime value using machine learning.

Wrapped in Kedro for deployment.

Overview

Can we build an accurate machine learning model to predict next 6 month LTV at customer level?

Analysis

• Please see ./ltv_model/model/notebooks/main.ipynb
• Steps to interact with notebook:
  • git clone https://github.com/jacobweiss2305/ltv_model.git
  • cd ltv_model
  • python -m venv venv (see provision for virtualenv installation)
  • pip install -r requirements.txt
  • source venv/bin/activate
  • If you want to run Causal Graphical Model and Bayesian Network:
    • Please see installation guide for pygraphiz: https://pygraphviz.github.io/documentation/stable/install.html
  • To spin up Jupyter lab you must use kedro:

    cd ./ltv_model/model/
    
    kedro jupyter lab
    

Provision

• I recommend using pyenv for python version control. We are running 3.8.9. Please see provision for details (model/docs/provision.md).