king-ak / spotify-charts

Example data product based around Spotify Chart data which leverages a medallion architecture

spotify-charts

Overview

This project is an example of a data product leveraging a medallion lakehouse architecture with Spark Streaming.

Dataset: Spotify Charts Dataset descripition from Kaggle:

This is a complete dataset of all the "Top 200" and "Viral 50" charts published globally by Spotify. Spotify publishes a new chart every 2-3 days. This is its entire collection since January 1, 2019. This dataset is a continuation of the Kaggle Dataset: Spotify Charts but contains 29 rows for each row that was populated using the Spotify API.

Gradle

Run tests with the following command:

gradle test

Build the project with the following command:

gradle build

Terraform Deploy

The pipeline can be deployed to Azure using Terraform with the following commands after the project has been set up with the databricks_poc repo

cd terraform
terraform init
terraform apply -auto-approve -var-file=environment/poc.tfvars \
  -var='databricks_workspace_resource_id=<DATABRICKS-WORKSPACE-RESOURCE-ID>' \
  -var='databricks_workspace_url=<DATABRICKS-WORKSPACE-URL>' \
  -var='poc_storage_account_key=<STORAGE-ACCOUNT-KEY>'

Terraform Destroy

The pipeline can be destroyed with the following command:

terraform destroy -auto-approve -var-file=environment/poc.tfvars \
  -var='databricks_workspace_resource_id=<DATABRICKS-WORKSPACE-RESOURCE-ID>' \
  -var='databricks_workspace_url=<DATABRICKS-WORKSPACE-URL>' \
  -var='poc_storage_account_key=<STORAGE-ACCOUNT-KEY>'

Local Run

The pipeline can be run locally with spark-submit (spark version >=3.5) using the following command:

bash utils/local-run.sh