A possible alternative is using the geom_dotplot function. In the case of an overlap, it jitters the overlapping points just enough to avoid them intersecting, while non-overlapping points are left untouched. See this link for more:

http://www.sthda.com/english/wiki/ggplot2-dot-plot-quick-start-guide-r-software-and-data-visualization

A workaround could be create a dummy variable in your data set that indicates overlap. Then during the graph creation use two geom_point and in one of them apply alpha

Ali Ezzat Thanks for your suggestion. My personal preference is to avoid jittering/dodging whenever possible–they often create visually appealing graphs at the cost of data misinterpretation.

Moreno S. Rodrigues Thanks. For a large dataset without creating a dummy "overlap" variable (which can be used to determine required degree of transparency), it's laborious/inefficient to manually create a vector of transparency for each data point. I can probably write a sub-routine that'll calculate overlap of a data point in a given window, but wondering if you know any package/function already does that?

Riccardo Esclapon

You're welcome. One last suggestion from my side would be:

If points are overlapping, but they are not exactly on top of each other, try adding a slim border to each point so that the user may be able to tell at a glance how many points are overlapping. You may try doing this with and without transparency.

Otherwise, consider using a shape other than circles (maybe triangles or "+" signs would be easier to view when overlapping?)

Hope this helps.