Scatter Plots

A scatter plot shows pairs of values as points.

Use it when you want to inspect a relationship without drawing a connected line:

• feature versus target;
• prediction versus true value;
• two measurements for each example;
• clusters or outliers.

plt.scatter(x, y)

Plot paired measurements

The plot suggests that larger hours usually goes with larger score in this tiny dataset. It does not prove a general law. It is an inspection.

Prediction versus target

A common ML scatter plot is:

plt.scatter(target, prediction)

If predictions are good, points often lie near the diagonal line y = x. That diagonal can be drawn as a reference.

plt.plot([0, 10], [0, 10])

Watch for swapped arrays

Scatter plots can reveal a pairing mistake. If the pattern disappears after sorting or shuffling one array, the pairs may no longer match.