SwapSmart Analytics

SwapSmart Analytics is a comprehensive end-to-end DBT Core analytics solution designed for EV companies focused on battery swapping in East Africa. This project aims to optimize operational efficiency and enhance customer experience by leveraging data-driven insights, specifically for the battery swapping model where clients exchange depleted batteries for fully charged ones.

Key Technologies and Tools

• Apache Airflow (Astro): For orchestrating data pipelines
• Soda: For ensuring data quality
• DBT: For transforming and modeling data
• Python: For generating synthetic data
• Metabase: For visualizing insights

Visualizations

Airflow Dags

Soda Checks

Data Modelling

Getting Started

1. Clone the repository:

   git clone https://github.com/Afrawles/SwapSmart-Analytics.git

2. Navigate into the project directory:

   cd SwapSmart-Analytics

Prerequisites

1. Rename the example configuration file:

   mv include/soda/configuration.yaml.example include/soda/configuration.yaml

2. Set up Soda account via Soda Cloud to monitor quality checks in the pipeline.

   • Click on your profile and create API Keys.
   • Paste the output into include/soda/configuration.yaml.

3. Start Astro and Airflow:

   astro dev start

4. Open a bash shell in the Astro container:

   astro dev bash

5. Navigate to the DBT directory, install dependencies, and exit:

   cd include/dbt/
   dbt deps
   exit

6. Access Airflow UI at: http://localhost:8080

   • Setup Postgres Variables Admin > Variables > Add New Record then fill form > Save
   • Run the generate_data DAG
   • Run the dw_bt DAG

7. Access Metabase at: http://localhost:3000 or use the IP found in Docker:

   • Inspect the container for the IP address using docker inspect <container-name> and look under Networks to get the IPAddress.

8. Set up DB connection in Metabase and build dashboards using the report tables.

Resources

• YouTube Channel: Marc Lamberti
• Connecting Metabase to PostgreSQL - 2 Docker Containers: Metabase Forum

Contributing

Contributions to improve the data model, add analytical queries, or enhance workflows are welcome. Please fork the repository, make your changes, and submit a pull request.

License

Contact

For questions or feedback regarding SwapSmart Analytics, please contact Odeke at [email protected].

• LinkedIn
• GitLab Profile
• Portfolio