Skip to content

Amansoni045/Slip

 
 

Repository files navigation

title Slip - Churn Intelligence Platform
emoji 📉
colorFrom blue
colorTo indigo
sdk docker
pinned false

Slip — Churn Intelligence Platform

From Predictive Analytics to Strategic Intervention

Live Demo Streamlit App

Slip is a production-grade churn intelligence platform. It bridges the gap between historical data analysis and real-world customer retention by combining classical Machine Learning with modern Agentic AI reasoning.

🌐 Live Experience

You can access the hosted version of the platform directly:

Project Vision

Customer churn is more than just a lost subscription; it’s a disruption in growth. This project tackles the churn problem in two distinct phases:

  1. Phase 1 (Predictive): We built a robust Random Forest pipeline to identify at-risk customers with 84% accuracy.
  2. Phase 2 (Agentic): We developed a "Strategist" Agent using LangGraph and RAG to reason through a customer's specific pain points and generate professional, data-backed retention plans.

Core Intelligence Phases

Phase 1: Machine Learning Core

  • Analytical Verdicts: Real-time inference on customer profiles.
  • Model Transparency: Feature importance and performance metrics built-in.
  • Data Balance: Utilizes SMOTE to ensure minority churn signals are never ignored.

Phase 2: Agentic Retention Layer

  • Cognitive Flow: Built on LangGraph for multi-step "Chain of Thought" reasoning.
  • Dynamic Retrieval: Queries an internal knowledge base of playbooks via FAISS.
  • Interactive Persona: Accepts custom user instructions to refine the strategic focus.

🚀 Launching the Platform

Get Slip running on your local machine in seconds:

# 1. Setup your environment
python3 -m venv .venv && source .venv/bin/activate

# 2. Synchronize dependencies
pip install -r requirements.txt

# 3. Launch the Intelligence Center
streamlit run app.py

📖 Exploratory Documentation


☁️ Hugging Face Deployment & Usage

Slip is optimized for deployment on Hugging Face Spaces using Docker.

Cloning from Hugging Face

To clone the source code directly from the Hugging Face Space:

# Note: You will need Git LFS installed to pull the model artifacts
git clone https://huggingface.co/spaces/Manku69/slip-churn-intelligence
cd slip-churn-intelligence
git lfs pull

Deployment Notes

  • SDK: Docker (Debian Slim)
  • Large Files: Managed via Git LFS (tracking .pkl and .csv files).
  • Environment Secrets: Requires GOOGLE_API_KEY, GROQ_API_KEY, and MISTRAL_API_KEY to be set in the Space settings.

Team & Development

Built by the Slip core team. Visit our Contributors Guide to see the faces behind the code.

About

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

No contributors

Languages

  • Jupyter Notebook 64.8%
  • Python 17.9%
  • TypeScript 16.8%
  • Other 0.5%