EcomForecast is an intelligent forecasting platform designed to empower e-commerce stakeholders with advanced time series predictions. It integrates traditional statistical models and deep learning approaches (LSTM, GRU) to generate accurate sales forecasts from historical data.
Qualified as a finalist for ICCSI-2023 (CREST, IIT Madras) for innovation in business analytics and AI-driven decision support systems.
Enable e-commerce businesses to:
- Predict future sales trends
- Identify seasonal patterns and growth anomalies
- Make inventory, marketing, and financial decisions proactively
- 🧠 Multi-Model Support: Includes SMA, EMA, WMA, LSTM, GRU
- 🧾 CSV Upload Interface: Clean preprocessing pipeline to handle raw e-commerce sales data
- 📊 Visual Forecasting: Interactive plots showing actual vs predicted sales
- 🛠️ Modular Design: Easily plug in new models or switch strategies
- 🏆 Pitch Finalist: Selected for final presentation at IIT Madras (ICCSI-2023)
| Component | Technology |
|---|---|
| Forecasting Models | Pandas, NumPy, statsmodels, PyTorch |
| Deep Learning Models | LSTM, GRU (using PyTorch/Keras) |
| UI Interface (Optional) | Streamlit / Gradio |
| Visualization | Matplotlib, Plotly |
| Data Format | CSV time series |