Choose a stage. Tick the checklist as you go; your ticks stay in this browser only. Stages 3 to 7 repeat every cycle; stage 8 happens at least once a year.
The same eight stages work from the command line. These commands make a folder of templates and a checklist, then say which stages have evidence:
python -m dqkit playbook init births --asset births-register --title "Births register"
python -m dqkit suggest data.csv --id births-register --out births/births-register.rules.yaml
python -m dqkit check data.csv births/births-register.rules.yaml --out births/results --history births/history
python -m dqkit playbook status births
AI assistants connected to the toolkit's MCP server can use the playbook tool for the same answers.
| In this playbook | You might call it |
|---|
R = does the work, A = accountable (one per stage), C = consulted, I = informed.
| When | What | Who |
|---|
| Level | Trigger | Action | By |
|---|
| Measure | Target |
|---|
Each downloads as a CSV file with the column headings, ready for a spreadsheet. Use roles and contact forms, not personal names or emails.
Poor data quality often falls hardest on people already under-served. Ask these at every stage.