FLIGHT DELAYS

I prepared the data, configured the dataset, data types and calculated fields, then built a map, line and bar visualisations connected by an airline selector. The dashboard itself didn't need SQL or Python — stated plainly. Separately, on top of the same domain, I built a reproducible SQL artefact: a three-table schema with primary and foreign keys, and eight queries — joins, aggregates, CASE, CTEs, window functions, an anti-join, a data-quality check. Every query is checked by a script that runs them in SQLite against an expected result.