Welcome back to our Sunday hands-on build! Modern analytics workflows often break down when data is scattered across data warehouses, operational databases, and unstructured cloud storage. In this practical session, we'll bridge that divide using Google’s Data Agent Kit in the Antigravity IDE to orchestrate, clean, test, and forecast data with intelligent agent tooling.

You will see firsthand how an agentic workflow streamlines complex data operations—from finding hidden assets to self-healing SQL transformations and generating in-engine forecasts.

🏗️ Today’s Technical Roadmap

We will step through a complete end-to-end data intelligence workflow:

1. Unified Asset Discovery: Map out and catalog cross-cloud data assets across BigQuery, Cloud SQL, and Cloud Storage (GCS) using the Dataplex Knowledge Catalog directly inside the Antigravity workspace.

2. Federated Anomaly Investigation via MCP: Query and correlate transactional records and analytical tables simultaneously in a single conversational session using Model Context Protocol (MCP) tools.

3. Automated dbt Transformation Pipelines: Prompt the agent to generate modular dbt (data build tool) staging models, write automated schema tests, and cleanly join multi-service datasets.

4. Self-Healing SQL & Fan-Out Debugging: Walk through a live debugging scenario where the agent identifies join cardinality flaws, diagnoses a data fan-out bug, and refactors the SQL models autonomously.

5. Predictive Analytics with BigQuery AI.FORECAST: Run built-in time-series projections directly inside BigQuery using AI.FORECAST to output actionable, forward-looking business recommendations.

🌟 Why This Session is Critical

• Agentic Data Orchestration: Learn how to use MCP interfaces to let AI agents interact safely and directly with multiple database engines.

• Resilient Data Pipelines: Combine automated agent reasoning with dbt best practices to build self-healing, well-tested data models that prevent data quality issues.

• In-Warehouse Machine Learning: Master BigQuery's native SQL-based predictive tools without having to export datasets to external Python ML environments.

• Production Portfolio Asset: Building a multi-service data discovery, transformation, and forecasting pipeline is a high-impact proof-of-work project that highlights advanced cloud and data engineering skills.

📝 What You Need to Bring

• Your Laptop (fully charged).

• Antigravity IDE Installed: Ready in your local development workspace.

• A Google Cloud Project: With billing enabled for BigQuery, Cloud SQL, Cloud Storage, and Dataplex.

• A GitHub Account: Ready to commit and version-control your dbt models and SQL scripts.

📚 Essential Developer Resources

• BigQuery Documentation: Forecasting with AI.FORECAST

• Dataplex Catalog: Data Discovery and Governance on GCP

• dbt Cloud / Core: Getting Started with dbt and BigQuery