What are we counting?
A sign-up is not a person. An arrival is not a session. Name the thing before you compare the numbers.
A FIELD GUIDE TO BETTER QUESTIONS
CONTEXT BEFORE CONCLUSIONSA little context. A better angle. A way to see what the numbers are actually saying. Explore a question-led approach to data, one small sample at a time.
Explore the sampleWhat did people come to learn?
ILLUSTRATIVE SAMPLE DATABEFORE THE ANSWER
Give the data a frame. Name what is included, notice what is missing, and leave enough room for the next question.
A sign-up is not a person. An arrival is not a session. Name the thing before you compare the numbers.
A single day and a four-week programme tell different stories. Keep the time window beside the result.
A chart can show a difference. It cannot supply a reason that the original records never captured.
THE SAMPLE FIELD
Three questions from an imagined workshop programme. Choose an angle. Read the shape. Inspect the rows.
Sample data only · no live AI or database connection
ANATOMY OF A USEFUL RESULT
Something you can read, question, and trace back to its source.
“What is happening?” is a start. “Which workshop had the most sign-ups in this programme?” gives the result a useful boundary.
Keep the unit visible. These samples distinguish sign-ups, arrivals and people, because those are different things.
Switch to the table to inspect every label and value. The drawing and the table use exactly the same source rows.
A useful answer says what it shows, and where it stops. The next question is part of the result.
QUESTIONS WORTH BRINGING
A better prompt starts before the prompt box. These are useful places to begin.
Start with a named measure and comparable windows. Ask whether the way it was recorded changed, too.
Look beyond a single total. A distribution can show differences that the average leaves out.
Notice whose experience never made it into the rows. Missing records are not the same as zero.
Choose one of the three authored questions. Read its context before interpreting the shape.
Move between chart and table. Follow the same values through a different way of reading.
Write a short local note about what you would want to know next. It stays on this page, unsent.
KEEP THE NEXT QUESTION
Turn a small observation into a question worth following. Your current sample choices are already here.
A local, unsent note.
Fictional concept. No AI execution, account, upload or submission. Use invented details only.
A LITTLE MORE CONTEXT
No. QUERY FIELD is a fictional product concept with three authored sample datasets. The questions select existing examples; nothing runs against a live AI model, database or analytics account.
They were invented for an imaginary workshop programme to demonstrate different ways of reading data. They are not measurements from a real organisation, customer or study.
Nowhere. Preparing a note creates an unsent summary in this page’s memory. There is no storage or submission. Reloading or resetting clears it. Please use invented details, with no personal or confidential information.
No. A chart and its table use the same labels, values and unit. Switching the question selects a different sample with its own context, measure and chart type.
Yes. Use Edit note or change any field or sample choice. The previous receipt clears, so you can prepare a new summary of your current thought.