SmartStockAI is an AI-driven solution that optimizes inventory management by predicting inventory trends, identifying potential deadstock risks, and generating actionable insights to minimize losses. Leveraging advanced machine learning models and large language models (LLMs), it transforms traditional inventory management into a dynamic, data-driven process adaptable to changing market conditions and consumer behaviors.
🏆 Winner of Google Build and Blog Marathon '24
🎥 Watch Demo Video
🚀 Deployment: Initially deployed on Cloud Run, but due to cloud charges, it has been removed. All functionalities are showcased in the demo video.
📝 Read the Full Story: Medium Blog
✅ Demand Forecasting - Uses the ARIMA_PLUS model in BigQuery ML to predict future sales and identify potential deadstock.
✅ Unstructured Data Analysis - Utilizes Gemini 2.0 LLM to extract insights from customer feedback and vendor notes.
✅ Actionable Recommendations - Generates strategies to reduce waste and improve efficiency in inventory management.
✅ Interactive Visualization - Provides dashboards via Streamlit and Looker Studio to visualize forecasts and insights.
SmartStockAI integrates multiple Google Cloud services for a seamless, scalable solution:
- BigQuery ML - Implements the ARIMA_PLUS model for demand forecasting.
- Gemini 2.0 LLM - Processes unstructured data to generate insights.
- Cloud SQL - Stores structured relational data.
- BigQuery - Serves as the core analytics engine for large-scale data processing.
- Streamlit - Provides an interactive frontend for data visualization.
- Looker Studio - Offers collaborative dashboards for deeper analysis.
Before implementing SmartStockAI, ensure you have the following:
- Cloud Storage
- Cloud SQL
- BigQuery
- BigQuery ML
- Looker Studio
- ARIMA_PLUS Model (for demand forecasting)
- Gemini 2.0 API (for unstructured data analysis)
- Streamlit (for visualization)
- SQL Queries
- Machine Learning Concepts
- Python Programming
- Google Cloud Platform (GCP)
Follow these steps to set up and run SmartStockAI:
1️⃣ Data Acquisition - Obtain inventory data (e.g., the Nike Sales dataset from Kaggle).
2️⃣ Data Upload - Upload the dataset to Google Cloud Storage.
3️⃣ BigQuery Integration - Enable BigQuery and connect it to Cloud SQL for real-time data retrieval.
4️⃣ Model Implementation - Apply the ARIMA_PLUS model in BigQuery ML for demand forecasting.
5️⃣ Unstructured Data Processing - Integrate Gemini 2.0 LLM for analyzing customer feedback and vendor notes.
6️⃣ Visualization - Develop interactive dashboards using Streamlit and Looker Studio.
🔗 BigQuery ML ARIMA_PLUS Model
🔗 Google Cloud Storage Documentation
🔗 Looker Studio Documentation
A huge thanks to Code Vipassana for organizing the in-person event! 🎉
This project is licensed under the MIT License. See the LICENSE file for details.
Feel free to open an issue or reach out via [LinkedIn/Twitter/GitHub Discussions]!
⭐ If you find this project useful, don't forget to give it a star! ⭐