Data-story reversal

Given a finished chart, reconstruct likely raw data shape and the claim it was built to support: reading graphics backwards.

When to use this: When you want to go deeper
Learning impact: Reliable gain
Learning evidence: Credible with caveats
Where you are with the topic: Applying in new contexts
How often to use it: Worth repeating often

When to use it

You are looking at a finished chart and want to read it backwards: data and claim.
You can redesign colours but cannot state what claim is being sold.

What to ask AI

Type something like: “Chart description: [ ]. I will reverse: inferred data structure + sold claim + what would falsify. Critique over-inference. Do not tell me the ‘right’ claim first.”

Ask it to critique your reversal, not to announce the claim before you try.

Keep the hard part on your side: you state data shape, claim, and falsifier first.

How it helps

Charts smuggle arguments. Reversing asks what observations would produce this picture, and what claim is being sold.

That builds skepticism and pairs with catching deceptive design.

How to run it

Take a chart from news, a paper, or a textbook.
Infer variables, grain, and what might be missing.
State the intended claim.
List what data could overturn it.
Compare to the caption and note spin.

Example exchange

You

There's a line chart labelled engagement that climbs to the right. I'm going to reverse-engineer the data and the claim. Please don't tell me the claim first.

What do you think each point is: daily totals, averages, or a selected cohort?

You

Looks like monthly active users. Claim: the product is growing. That would be false if the y-axis is a tiny truncated range or bots were counted.

Solid start, but "users" might be over-inferred. It could be clicks. What caption words would confirm the grain before you trust the growth claim?

Copyable prompt

Chart description/data ink: [ ].
I will reverse: inferred data structure + sold claim + what would falsify.
Critique over-inference. Do not tell me the "right" claim first.

The Tell

Here is how you know this method has flipped: you can redesign colours and still cannot state the claim or a falsifier.

Trusting the caption blindly also skips the method.

Leave with claim + falsifier in your own words.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Chart → inferred data + claim is critical graph literacy run backwards: visualisations as argumentative commitments. It pairs with chart deception analysis.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI reverse-engineering of charts is speculative and can over-infer. Your explicit uncertainty about grain and missingness is part of a good reversal.