Columnar stores like ClickHouse enable users to pull insights from big data in seconds, but only if you set things up correctly. This talk will walk through how to implement a data warehouse that contains 1.3 billion rows using the famous NY Yellow Cab ride data. We'll start with basic data implementation including clustering and table definitions, then show how to load efficiently. Next, we'll discuss important features like dictionaries and materialized views, and how they improve query efficiency. We'll end by demonstrating typical queries to illustrate the kind of inferences you can draw rapidly from a well-designed data warehouse. It should be enough to get you started--the next billion rows is up to you!