https://github.com/jra333/causal-lift-analysis

A causal lift analysis demonstrating the impact of certain metrics have on sales + POC app usecase.
https://github.com/jra333/causal-lift-analysis

analytics causal-inference predictive-modeling regression streamlit structual-equation-modelling

Last synced: 7 months ago
JSON representation

A causal lift analysis demonstrating the impact of certain metrics have on sales + POC app usecase.

• Host: GitHub
• URL: https://github.com/jra333/causal-lift-analysis
• Owner: jra333
• Created: 2025-02-06T08:10:35.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2025-02-07T02:56:51.000Z (8 months ago)
• Last Synced: 2025-02-07T03:27:09.928Z (8 months ago)
• Topics: analytics, causal-inference, predictive-modeling, regression, streamlit, structual-equation-modelling
• Language: Jupyter Notebook
• Homepage:
• Size: 702 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Structural Equation Modeling (SEM) Analysis & App

**Two folders include:**

- SEM [App](https://github.com/jra333/sem_causal_lift/tree/main/sem_app_testing) POC using Streamlit

An interactive web application built with [Streamlit](https://streamlit.io/) that enables users to perform Structural Equation Modeling (SEM) on their data. The app integrates data upload, preprocessing (with time series support), dynamic model specification, SEM analysis using [semopy](https://github.com/semopy/semopy), and AI-generated interpretations via Google Gemini.

- Causal Impact Analysis [Notebook](https://github.com/jra333/sem_causal_lift/tree/main/sem_notebooks)

The integrated impact analysis was able to prove that upper-funnel metrics contribute to sales through modeling and predicting sales using a scenario based approach. Explored and confirmed the lift in sales using a custom structural equation model highlighting the pathways in upper-funnel that lead to sales increase.

  - Performs comprehensive data aggregation (daily, weekly, and monthly) on raw sales data.

  - Implements advanced visualization techniques including time series plots, correlation heatmaps, and channel comparison plots.

  - Conducts integrated causal impact analysis combining Random Forest and OLS regression approaches.

  - Provides detailed elasticity analysis with statistical inferences and seasonal impact assessments.