dive

the scatter · do crowded places have worse air?

The same data, drawn four ways.

Real datasets with the chart already drawn. You read it and decide what it shows. The partner helps along the way.

start a dive how it works

how it works

Guess first, then look. The claim is yours.

  1. 1

    guess

    Lock in your guess before you see any chart. You cannot change it later.

  2. 2

    look

    Read the chart: the axes, the units, and what each mark stands for. Write down what you found.

  3. 3

    claim

    Test it on the same data. Then write your claim: what you think, why, and how you checked.

  4. 4

    check

    Now see the numbers. Was your guess right? If not, say what was different.

the partner

It asks. You read.

The partner is Claude, an AI. It asks what you see first, points to the one thing worth a second look, and names the idea so you know the word next time. It never writes your claim.

youso crowded counties have worse air?
the partnercorrelationLook at Los Angeles and San Bernardino: close on air, far apart on crowding. What would you need to see before you said "because"?

your fluency

Working with AI is a skill. dive tracks how yours grows.

iterate

You build on the partner's last answer instead of starting over.

clarify

You ask with detail: which measure, which years, which group.

approach

You ask how to read a chart before asking what it means.

context

You tell the partner when it got you wrong or missed something.

These come from the four Ds, the same record spawn, bench and claim keep. dive also counts your guesses and the times you explained a miss. It keeps counts only, never your words. You and your advisor see them on the growth pages. None of it is a grade, and none of it changes what you can do.