overplotting

Overplotting is a visualization problem that occurs when a dataset contains many observations with similar values, causing marks to overlap heavily and obscure the true structure of the data. It is particularly common in scatter plots where points are plotted at precise coordinates, and in plots of variables with limited granularity or high cardinality.

Causes: High data density relative to the plot resolution; overlapping values across observations; large sample sizes; use of markers with default size; identical or near-identical coordinates; low variability.

Effects: Dense regions appear as solid blobs, making it hard to discern density, clusters, trends, or correlations; outliers may be hidden; apparent relationships can be exaggerated or masked by piling of points; color mixing can distort interpretation when many points share colors.

Mitigation: Increase transparency (alpha) or reduce point size; add jitter to separate points; use hexbin plots or 2D histograms to aggregate data; apply kernel density estimates or contour density plots; use marginal plots, faceting, or small multiples; sample the data or plot a subset; prefer interactive plots that allow zooming and brushing.

Contexts: Overplotting is a common consideration in exploratory data analysis, especially with large datasets, high-resolution displays, or data with discretized values; awareness of overplotting guides choices of visualization techniques to reveal underlying patterns.