Does r = 0 mean the variables have no relationship at all?
It means there is no detectable linear relationship between them; two variables can have a strong but curved relationship (like y = x²) and still produce an r near zero, because the Pearson coefficient only picks up the linear component of the association.
Can a single outlier change the correlation's classification?
Yes, and it is one of the most common risks: since r uses the sum of deviations from the mean, one point far removed from the rest can pull r from moderate to strong (or from weak to moderate) on its own, without the pattern in the rest of the data having changed.
What is the difference between correlation and causation?
Correlation measures statistical association, not mechanism; two variables can correlate by coincidence, through a third variable that influences both (a confounding variable), or through real causation, and the Pearson coefficient alone cannot tell these three possibilities apart.
Why does the tool require at least 2 points and at most 200?
With fewer than 2 points there is no way to compute a deviation from a statistically meaningful mean; the 200-point ceiling exists to keep manual entry practical, since the form accepts typed or pasted rows, not an external data file.