Bug. In Prism 8.0.0 and 8.0.1 points in a scatterplot can nearly overlap so you don't really see all the data

Problem

Scatter graphs plot every value, moving points left or right to avoid overlap. In Prism 8.0, we improved and revised the algorithm for deciding how far each symbol is from the center of a scatter plot. We updated the algorithm that previous versions of Prism used that sometimes ended up arranging the points in a way that created "smiles" in the data. The new algorithm in Prism 8.0 eliminated this patterning of the data, so scatter plots provide a better representation of the overall distribution of the data. 

However, sometimes the algorithm in 8.0.0 and 8.0.1 doesn’t work perfectly, and points can be nearly superimposed. This is a problem because the whole point of a scatterplot is to clearly show every point. 

Workarounds

In Prism 8.2.0, we introduced the ability to choose between three different algorithms that control how the data points in a scatterplot are arranged. These options are:

• Standard: the width of the distribution of points is proportionate to the number of points at that Y value. This is the best option to represent the overall distribution of the data
• Classic: this algorithm prioritizes minimizing the amount of overlap of individual data points over representing the shape of the data distribution. This algorithm may also cause "smiles" in the scatterplot
• Expanded: this algorithm also prioritizes minimizing the amount of overlap of individual data points over representing the shape of the data distribution. However, this algorithm also attempts to remove any visual patterns in the data

Note that depending on the amount of data, the shape and size of the symbols, and the size and spacing of the graph, there may be no way to completely avoid overlapping symbols. Prism moves symbols left and right when attempting to avoid overlap of symbols, but will always plot the symbol at its given Y value. With enough data points at the same (or very similar) Y value(s), there may simply not be enough space to place every single data point without overlap. In this situation, there are additional alternative workarounds that may be used to try and make points better stand out from each other:

• Choose a semitransparent color for symbols so points get darker when symbols overlap
• Use smaller symbols
• Make the graph larger by selecting the Y-axis and dragging the knob
• Select a different symbol shape (try diamonds)
• If you have lots of points, plot a violin plot rather than a scatter plot