Five pieces: AI, trust and governance in the maritime industry, a count of how much open AIS data is personal data, a Databricks funnel that names the leak, a Snowflake cleaning pipeline that turns messy exports into an analysis view, and what that cleaning changes on a dashboard.
GDPR, the 6Ws, and a hazard taxonomy for AI accidents. Where governance, trust and ethics meet in a fragmented, safety-critical industry.
170 million AIS messages from Danish waters. Two thirds of the vessels are private boats — but only a quarter of the data, and the share collapses in winter.
Databricks Lakehouse · 50,000 sessions · Medallion architecture. Where shoppers drop off, and why they leave.
Snowflake · Orders & Products. Five price formats, five date formats, flagged rather than deleted, then joined.
The same 5,000 orders charted twice: raw, the way an AI dashboard reads them, and cleaned in Power BI. The raw view misses 69% of revenue, and buries the refunds.