Scatterplot

A scatterplot is a two-dimensional data visualization used to examine the relationship between two quantitative variables. Each observation is represented by a single point with coordinates corresponding to the values of the two variables. The pattern formed by the points can indicate the strength, direction, and form of any association between the variables.

Construction and interpretation

Axes are labeled with the respective variable names and units, and scales are chosen to reflect the data range. Points may be transparent or slightly jittered to reduce overplotting when many observations share similar values. A regression line or smoothing curve can be added to summarize the overall trend, but such additions do not imply causation. The scatterplot is commonly used in exploratory data analysis to identify linear or non-linear relationships, clusters, outliers, and potential data quality issues. The strength of association can be assessed quantitatively with correlation metrics or locally with nonparametric fits.

Variants and enhancements

A scatterplot matrix (or pairs plot) extends the idea to multiple variables by displaying all pairwise scatterplots in a grid. Three-dimensional scatter plots visualize a third variable using depth, while hexbin plots or transparency-based encodings help with large datasets. Color, shape, or size can encode additional categories or measurements to convey more information within the same plot.

Limitations and best practices

Scatterplots show association but not causation and can mislead if scales are distorted or if subsamples bias the view. Careful labeling, consideration of outliers, and awareness of confounders are important for valid interpretation.