End-to-end data transformation pipeline built with dbt, BigQuery, and Python. This project simulates a real-world logistics scenario, processing 50,000+ trip records to analyze fuel efficiency across a commercial fleet.

The project is structured following the Medallion Architecture to ensure data traceability and quality:

1. BRONZE Layer (Staging)

• stg_trips: Technical cleansing, date normalization, and sensor error filtering.
• stg_vehicles: Standardization of truck fleet metadata.
• stg_drivers: Processing of driver master records.

2. SILVER Layer (Intermediate)

• silver_fleet_performance: Integration table joining telemetry (trips) with dimensions (drivers and vehicles). Includes fuel efficiency business logic and outlier handling.

3. GOLD Layer (Marts)

• gold_fleet_stats: Final reporting table for business stakeholders. Contains aggregated metrics and performance rankings by model and driver.

Robustness is guaranteed through dbt tests:

• Generic Tests: not_null and unique on primary keys.
• Business Tests: dbt_utils.accepted_range to ensure fuel consumption and distances fall within realistic physical bounds (e.g., 0 to 200 L/100km).

• Data Transformation: dbt (Data Build Tool)
• Warehouse: Google BigQuery
• Environment: Conda
• Data Generation: Python (Pandas/Numpy)
• Visualization: Looker Studio

This project utilizes the following dbt packages to extend functionality:

• dbt-utils: Used for advanced data quality testing (accepted_range) and cross-database macros.

1. Clone the repo.
2. Setup your profiles.yml for BigQuery.
3. Install dependencies: dbt deps.
4. Run the pipeline: dbt run.
5. Execute tests: dbt test.