YouTube video by Microsoft Hates Greg

YouTube video by Joe Freeman: Software Craft, Org Stuff, Tech Stuff

Hi For years, organisations built data warehouses because they were considered “standard infrastructure.” Now we’re doing the same thing with data lakes. The problem is not the technology. It never was. The problem is that most organisations are building storage before they build accountability, governance, ownership, or operational discipline. The result is predictable: platforms full of data nobody trusts, nobody understands, and nobody can confidently use for AI or decision-making. In my latest article, I explore: • Why 80% of data lake initiatives fail to deliver value • How organisations accidentally create “data swamps” • The governance mistake boards continue to underestimate • Why AI strategies fail when built on unmanaged data foundations • Three practical frameworks I use with organisations to change the outcome Including: → DAVE Reversed - start with the decision, not the data → VIA - Value, Intent, Accountability → The governance principle most organisations implement far too late One organisation I worked with invested heavily in a data lake for machine learning and AI use cases. But nobody had labelled the data. The platform technically worked. The business outcome did not. That story is becoming far more common than many leadership teams realise. If your organisation is investing in AI, cloud platforms, analytics modernisation, or enterprise data programmes, this is ultimately not a technology conversation. It is a leadership and governance conversation. Read the full article below. As always, I’d love to hear your perspective. Liz Henderson The Data Queen Board Advisor, Data & AI Strategy

Aurora PostgreSQL can now query Parquet and Iceberg data in S3 through an embedded DuckDB engine. What changes for data pipelines, and what does not.

前回は、Oracle AI Data Platform（AIDP）を作成し、外部Databaseの表へSQLを実行してみてみました。今回は、Databaseの外にあるデータをAIDPへ取り込み、SQLで比較できる形に整えるところから試してみます。 使用するのは、e-Sta...

Starting with a local Parquet file before joining it to data stored in Amazon S3