📌 Overview

Retail sales forecasting application using Prophet and ARIMA deployed via Streamlit. Enables interactive demand planning through real-time dashboard visualizations with configurable forecast horizons.

📊 Model Performance

Metric	Score
MAPE	8–12%
Models Used	Prophet, ARIMA
Deployment	Streamlit Web App

🔑 Key Features

• Dual-model forecasting: Prophet for trend + seasonality, ARIMA for stationary series
• Interactive UI: Streamlit dashboard with configurable date ranges and forecast windows
• Plotly visualizations: Trend decomposition, seasonality analysis, forecast confidence intervals
• Automated EDA: Summary stats, stationarity tests, ACF/PACF plots
• CSV upload: Works on any time-series dataset

🗂️ Project Structure

sales-forecasting-dashboard/
├── app.py                    # Main Streamlit application
├── src/
│   ├── forecaster.py         # Prophet & ARIMA logic
│   ├── visualizer.py         # Plotly chart generation
│   └── preprocessor.py       # Data cleaning & transforms
├── data/
│   └── sample_sales.csv
├── requirements.txt
└── README.md

⚙️ Setup & Run

git clone https://github.com/Sandy7217/sales-forecasting-dashboard.git
cd sales-forecasting-dashboard
pip install -r requirements.txt
streamlit run app.py

Then open http://localhost:8501 in your browser.

📦 Dependencies

streamlit
pandas
numpy
prophet
statsmodels
plotly
scikit-learn

👤 Author

Sandeep Rana — GitHub · Email