Scatter plot

A graph that displays paired numerical observations as points, revealing their distribution and possible relationships between variables.

Bivariate data: Data in which each observation has measurements on two variables. Each point represents one paired observation across the plot’s two variables.
Correlation: A statistical measure of the direction and strength of association between variables. A scatter plot makes positive, negative, and weak correlations visible.
Exploratory data analysis: The use of summaries and visualizations to inspect data before formal modeling. Scatter plots reveal unexpected structure before assumptions are chosen.
Line chart: A graph that connects ordered observations with line segments to emphasize sequence or continuity. Connecting points can imply order or continuity that paired observations may not have.
Cartesian coordinate system: A system that locates points using numerical coordinates on perpendicular axes. The horizontal and vertical coordinates position every plotted observation.
Linear regression: A method that models an outcome as a linear function of one or more predictors. A fitted line summarizes a roughly linear pattern among plotted points.
Anscombe's quartet: Four datasets with nearly identical summary statistics but strikingly different plotted patterns. Its plots show why numerical summaries cannot replace visual inspection.
Causal inference: The study of whether changes in one variable produce changes in another. Visible association alone does not establish that one plotted variable causes the other.
Independent variable: A variable treated as an input or predictor in an analysis. It is commonly assigned to the horizontal axis when displaying a relationship.
Least squares: A method for estimating model parameters by minimizing the sum of squared residuals. It commonly determines the regression line drawn through scatter-plot observations.

Linked from 10 pages