About the Project Inspiration
Global supply chains are becoming increasingly complex, data-heavy, and unpredictable. Yet, many organizations still rely on delayed or manual data processing. We wanted to harness the power of Fivetran and Google Cloud AI to build a unified, real-time analytics platform that transforms raw logistics data into actionable insights.
What it does
Our dashboard connects multiple data sources—procurement, warehousing, logistics, and sales—through Fivetran’s automated pipelines. It then visualizes this data in real time on Google BigQuery, powered by Vertex AI models for predictive forecasting, anomaly detection, and operational recommendations. Users can monitor KPIs like delivery efficiency, inventory levels, and transportation costs, with AI-driven alerts for potential disruptions.
How we built it
Fivetran Connector SDK to automate ETL pipelines from ERP, IoT, and CRM systems.
Google BigQuery for scalable data storage and querying.
Vertex AI for building predictive and anomaly detection models.
Streamlit + Looker Studio for dashboard visualization and interaction.
Cloud Functions to automate model retraining and alert triggers.
Hosted the entire stack on Google Cloud Run for scalability and accessibility.
Challenges we ran into
Integrating multiple data sources with varying formats and update frequencies.
Designing a clean, real-time visualization layer with low latency.
Balancing AI model performance with interpretability for business users.
Managing limited dataset access for testing during early prototyping.
Accomplishments that we’re proud of
Built a fully functional real-time analytics pipeline using Fivetran + Google Cloud.
Developed an AI model that improved forecasting accuracy by nearly 20%.
Designed an intuitive dashboard that converts complex data into actionable insights.
Created a scalable, modular architecture ready for enterprise adoption.
What we learned
Hands-on experience with Vertex AI pipelines, BigQuery optimization, and automated ETL workflows.
How real-time data can reshape supply chain decision-making.
The importance of designing for usability — not just accuracy.
What’s next for AI-Driven Supply Chain Analytics Dashboard
Integrate Gemini AI for conversational query capabilities.
Add RAG (Retrieval-Augmented Generation) for contextual data insights.
Enable simulation models for “what-if” supply chain scenarios.
Extend the dashboard to sustainability metrics like carbon tracking.
Built With
- cloud-functions
- cloud-run
- cloud-storage
- fivetran-connector-sdk
- gemini
- google-bigquery
- looker-studio
- python
- streamlit
- vertex-ai
Log in or sign up for Devpost to join the conversation.