A scatter plot shows how two variables move together. Each observation contributes one pair (x, y), drawn as a single dot: x sets how far right it sits, y how far up. Plot the whole sample and the cloud of dots reveals the relationship between the two variables at a glance.

One dot per observation — so a class of 30 students, each measured for height x and shoe size y, becomes 30 dots. Nothing is summarised away; you see every individual and the overall shape at the same time.

Direction and shape

Two questions read straight off the picture.

• Direction. If the dots drift upward to the right — large x with large y — the association is positive. If they drift downward — large x with small y — it is negative. A shapeless blob with no consistent tilt shows no association.
• Shape. If the dots hug a straight path the relationship is linear; if they follow a bend or a U the relationship is curved (nonlinear).

Switch between three clouds

Flip the control to compare a positive cloud (rising to the right), a negative cloud (falling to the right), and a cloud with no association (a shapeless scatter). The same axes hold all three, so only the tilt of the dots changes.

Association is not causation

A scatter shows that two variables move together — it does not show that one causes the other. Ice-cream sales and drowning deaths both climb in summer, so they form a tidy positive scatter, yet neither causes the other: a third, lurking variable — the heat — drives both. A pattern in the dots is a starting point for investigation, never a proof of cause.