The problem it answers
- Questions answered by anecdote
- Analysts rebuilding the same extract
- No shared definition of a metric
How this category works
Data work follows a lifecycle: sources, collection, storage, engineering, analytics, visualization, and the decision. Dashboards without a maintained pipeline become stale. Warehouses without a question become storage bills. We start from the decision.
1. Name the decisions
The reports and models follow the questions, not the other way around.
2. Make the data trustworthy
Pipelines, models, and a place the numbers agree.
3. Show it
Dashboards and reports for the people who act, with definitions written down.
4. Keep the pipe
Analytics and reporting are recurring. Sources change.
Questions
- We have reports already. Why a data project?
- When each team’s numbers disagree, or when the report is assembled by hand every week, the issue is the path of the data, not another chart.