pm-dataeng

Define a data contract between a producer and consumers of a dataset/event/API. Use when asked to write a data contract, define a schema agreement, set data SLAs, or stop a producer from silently breaking downstream consumers. Produces a contract — schema with types & constraints, semantics, quality SLAs (freshness/completeness/validity), ownership, versioning & breaking-change policy, and a change process.

Design the data quality checks for a table or pipeline across the standard dimensions. Use when asked to add data quality tests, define DQ checks, catch bad data before it hits dashboards, or set up monitoring for a dataset. Produces a checks plan across completeness, validity, uniqueness, freshness, consistency, and accuracy — each with the rule, severity, and where it runs (dbt test / Great Expectations / SQL assertion).

Spec a dbt model — its grain, sources, transformations, tests, and materialization. Use when asked to design a dbt model, plan a data transformation, write a staging/intermediate/mart model spec, or define dbt tests for a table. Produces a model spec — purpose & grain, lineage (sources → refs), the transformation logic, column definitions, dbt tests, materialization choice, and the skeleton SQL/YAML.

Analyse a finished A/B test and write an honest results readout with real statistics. Use when asked to read out an A/B test, analyse experiment results, check if a result is statistically significant, or decide ship/no-ship from test data. Produces a readout — the computed lift, p-value & confidence interval, a significance verdict, guardrail check, and a clear ship / no-ship / iterate recommendation. Includes a stdlib significance calculator.

Define a metric in a semantic layer so it means one thing everywhere. Use when asked to define a metric, build a semantic layer / metrics layer entry, stop 'revenue means three things' problems, or write a metric definition for dbt MetricFlow / Cube / LookML. Produces a metric definition — exact formula, the base measure & aggregation, dimensions, filters, grain, edge cases, and a tool-ready spec.

Diagnose a slow SQL query and produce a concrete optimization plan. Use when asked to optimize SQL, speed up a slow query, reduce a query's cost/scan, fix a timeout, or review a query plan. Produces an analysis — the likely bottleneck, what the plan is doing wrong (full scans, bad joins, spills), the specific rewrite and index/partition changes, and the expected impact, with the optimized query.