This blog post explores how to build and optimize a serverless data pipeline on AWS, transforming data from various sources into actionable business insights. Using Helsinki region public transport data as a real-world example, the article demonstrates how AWS Glue Jobs extract and transform data from DynamoDB tables into a centralized S3 data lake in Parquet format. AWS Glue Crawlers then catalog the data schemas, Amazon Athena provides SQL querying capabilities, and Amazon QuickSight delivers visualizations using its SPICE in-memory caching engine. The post details optimization strategies for AWS Glue ETL processes, including parallelization, data partitioning, and auto-scaling. Finally, it explains how AWS Step Functions and CloudWatch Event Triggering can automate the entire pipeline, ensuring QuickSight datasets refresh in the correct order immediately after Glue Workflows complete, delivering consistently up-to-date insights to business users.