Alexander Booth·Oct 12Test Your PySpark Like Real Software: pytest, Fixtures, and Databricks ConnectSeparate the math from the Spark, test both locally before they hit production, and run the exact same suite on your laptop, in a notebook…A response icon1A response icon1
Alexander Booth·Sep 28Streaming Without Kafka: Structured Streaming vs. Zerobus Ingest on DatabricksOne casino floor, 500 events, two architectures — and why, for single-destination flows, you can drop the message bus entirely.
Alexander Booth·Sep 14From a Music API to a Governed Data Product: The rostr Pipeline on DatabricksIngest an evolving music-industry API, run it through a medallion, and land a star schema plus a dashboard that answers “who represents…
Alexander Booth·Aug 31Every Event, One Star Schema: An Events Analytics Pipeline from the Ticketmaster APIPull a public events API, run it through a medallion, and land a governed star schema you can actually query — bronze to gold in five…
Alexander Booth·Aug 17Query Governed Delta Tables From Local Python: No Warehouse, No ClusterDelta Sharing is an open protocol. A partner can query your governed tables with DuckDB and pandas, from a laptop, with zero Databricks…A response icon1A response icon1
Alexander Booth·Aug 3Catch Schema Drift Before It Breaks You: VARIANT, a Drift Detector, and Alerts on DatabricksAn upstream API adds three fields. Your bronze layer shrugs, a detector lights up, an alert fires, and silver widens in place. No 2 a.m…
Alexander Booth·Jul 20From a Sports API to Self-Serve BI: A Governed Hockey Data Product on DatabricksIngest a hockey analytics API, land it through a medallion, model it as a star schema, and let a coach ask questions in plain English — all…
Alexander Booth·Jul 6Ask Your Stadium Concession Data in Plain English: Tuning Databricks GenieA venue ops manager shouldn’t need SQL to ask “how does rain affect beer sales?” With a well-tuned Genie space, they don’t.
Alexander Booth·Jun 22Expected Goals, End-to-End: An MLOps Pipeline for Hockey on DatabricksFrom synthetic shots to a calibrated xG model — governed in Unity Catalog, served behind an endpoint, and watched for drift.
Alexander Booth·Jun 4From Box Scores to Predicting the NBA Finals: An End-to-End NBA Model on DatabricksIngest, govern, model, and serve NBA data on Databricks.