By the end of the book BookNest has a small but complete data platform, and every chapter builds one part of it. The figure maps those parts onto the lifecycle, with the chapter that builds each.
Read the figure as a set of paths rather than one pipeline. Catalog XML is pulled as files, validated and transformed with XSLT, then stored in columnar files. Orders flow from PostgreSQL 1,289 through change data capture and Kafka 129 , land in Iceberg 129 tables, and are modeled with dbt 37,942 into the sales mart that analysts query. The same order events also feed a streaming job (the dashed path) that keeps live totals current. The Capstone assembles all of it into one capstone run, and A Tiny End-to-End Pipeline starts much smaller: one CSV file, one Parquet 129 file, one query and one chart.