Portable Airflow Data Stack

This application is an Analytics suite suite for an imaginary company selling postcards. The company sells both directly but also through resellers in the majority of European countries.

Stack

• Airflow
• Docker & Docker-Compose
• dbt core
• Superset

For other orchestration options, check the below:

• portable-data-stack-mage
• portable-data-stack-dagster

Setup

Rename .env.example file to .env and set your desired password. Remember to never commit files containing passwords or any other sensitive information.

System requirements

• Docker
• Docker-Compose

With Docker engine and Docker-Compose installed, change directory to the root folder of the project (also the one that contains docker-compose.yml) and run

docker-compose up --build

Demo Credentials

Demo credentials are set in the .env file mentioned above. For Airflow they are by default the following:

• AIRFLOW_USER=airflow
• AIRFLOW_PASSWORD=airflow

Ports exposed locally

• Airflow: 8080
• Superset: 8088
• PostgreSQL Data Warehouse instance: 54321
• PosgreSQL OLTP Database instance: 54320

Generated flat files (XML, CSV) are saved in the import folder.

The data is fictional and automatically generated. Any similarities with existing persons, entities, products or businesses are purely coincidental.

dbt core

If you'd like to run your dbt core model outside of Airflow, you can:

• create a python virtual environment

  python3 -m venv .venv

• activate it

  source .venv/bin/activate

• install the requirements found under dbt/requirements.txt

  pip install -r requirements.txt

• run dbt commands, like

  dbt compile

General flow

1. Generate test data (flat files + OLTP data) using Python
2. Import flat file data and OLTP data to staging area in the Data Warehouse, orchestrated by Airflow
3. Process data, build fact and dimension tables, load the Data Warehouse using dbt
   • installs dbt dependencies
   • seeds the database with static data (geography)
   • runs the model
   • tests the model
4. Analyze and visually explore the data using Superset or directly query the Data Warehouse database instance

For superset, the default credentials are: user = admin, password = admin

Overview of architecture

The docker-compose process will begin building the application suite. The suite is made up of the following components, each within its own docker container:

• generator: this is a collection of Python scripts that will generate, insert and export the example data
• oltp: this is the PostgreSQL instance that will simulate our transactional database (sales_oltp), serving as one of the sources of the data; this is locally available on the host machine exposed on port 54320.
• airflow: this is the orchestrator tool that will trigger the ETL tasks; its GUI is locally available on port 8080;
• airflowdb: this is a PosgreSQL instance, upon which airflow depends
• dw: this is a PostgreSQL instance that will host our Data Warehouse; locally available on port 54321 (database sales_dw).
• superset: this contains the web-based Business Intelligence application we will use to explore the data; exposed on port 8088.

Once the docker-compose building process has completed, we may open the Airflow GUI (locally: localhost:8080) to view the orchestration of our tasks. There are four defined workflows (known as DAGs in Airflow), as follows:

• initialize_etl_environment: this workflow initializes the ETL environment by creating the necessary database objects; this needs to run only once;
• import_main_data : this imports the data from the transactional system to a staging destination on the DW instance
• import_reseller_data: this workflow extracts, preprocesses and loads the XML and CSV data sent by the resellers into a staging destination on the DW istance
• run_dbt_init_tasks: this workflow initializes dbt (needs to only run once)
• run_dbt_model: this workflow transforms the data from transactional and flat-file sources, stores them in a staging area and loads the data warehouse, leveraging dbt-core

After the DAGs have completed you can either analyze the data using the querying and visualization tools provided by Superset (available locally on port 8088), or query the Data Warehouse (available locally on port 54321)