Customer Sales Forecasting and Churn Prediction

A full-stack data science project that forecasts customer sales and predicts churn using machine learning, visual analytics, and interactive dashboards. This project is containerized with Docker and production-ready with a REST API built using FastAPI.

Features

• Data Cleaning & Preprocessing
• Exploratory Data Analysis (EDA)
• Feature Engineering & Customer Segmentation
• Predictive Modeling (Sales Forecasting & Churn Prediction)
• Model Evaluation & Selection
• Interactive Dashboard using Streamlit
• REST API using FastAPI for model serving
• Dockerized for easy deployment

Project Phases

Phase 1: Data Cleaning & Preprocessing

• Handled missing values, outliers, and feature encoding
• Scaled features for modeling

Phase 2: Exploratory Data Analysis

• Visualized sales trends, customer segments, and patterns
• Grouped insights by product, region, and customer

Phase 3: Feature Engineering & Clustering

• Created new insightful features
• Applied KMeans clustering for segmentation
• Visualized clusters using heatmaps and elbow curves

Phase 4: Predictive Modeling

• Built and compared multiple models (RandomForest, XGBoost, etc.)
• Selected best model based on performance metrics
• Saved model with joblib

Phase 5: Streamlit Dashboard

• Built an interactive dashboard to visualize insights and make predictions
• Integrated trained model with Streamlit interface

Phase 6: API & Dockerization

• Created a FastAPI backend for real-time prediction
• Dockerized the API for scalable deployment

Tech Stack

Category	Tools/Technologies
Data Processing	Pandas, NumPy
Visualization	Matplotlib, Seaborn, Plotly
Machine Learning	Scikit-learn, XGBoost, KMeans
Dashboard	Streamlit
Backend API	FastAPI, Uvicorn
Model Serving	joblib
Containerization	Docker
Version Control	Git + GitHub

Installation & Usage

Requirements

• Python 3.10+
• Docker installed and running

Clone the Repository

Navigate to your desired folder and run:

git clone https://github.com/msaakaash/customer-sales-forecasting.git
cd customer-sales-forecasting

Optional: Create a Virtual Environment

• Windows:
  python -m venv venv && venv\Scripts\activate

• macOS/Linux:
  python3 -m venv venv && source venv/bin/activate

Install Dependencies (Development)

Install all required Python packages:

pip install -r api/requirements.txt

Run the Streamlit Dashboard

Launch the interactive dashboard:

streamlit run streamlit_app.py

Visit http://localhost:8501

🌐 Run the FastAPI Backend with Docker

cd api

docker build -t churn-api .

docker run -p 8000:8000 churn-api

Visit http://localhost:8000/docs for interactive Swagger docs.

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

Code of Conduct

Please read our Code of Conduct before contributing to this project.

Security

If you discover a vulnerability, please refer to our Security Policy for instructions on how to report it responsibly.

License

This project is licensed under the MIT License.

Author

Aakaash M S