[Field Guide]
Data Atlas
A structural map of a company's data operation — systems, tables, pipelines, metrics and traps — written for whoever arrives next.
What anyone knew about the data operation lived in my work logs, in chat threads and in people's heads. A new analyst had no way to know which table to trust.
Distilled about four months of measured findings from my work vault into a structured vault for the team.
Logs are organized by date. A newcomer needs answers organized by question: where is this number, how is it built, what breaks it.
01 · Structure, not diary
Sections for business, architecture, databases, pipelines, metrics and rules, dashboards, traps, runbooks and SQL recipes — one note per thing, updated in place.
02 · Every number is dated
Each figure says when it was measured, because the operation changes weekly.
03 · Traps as first-class content
Separate notes for date and timezone traps, money traps, grain and counting traps, and silent failures.
04 · A way in
A 10-minute overview, a first-week plan and a checklist to run before stating any number.
05 · Safe to share
No credentials or customer data; open questions are written as “to confirm with <role>”, never guessed.