Hey there!
I am currently incorporating a Dash app into my Django website and really appreciate the awesome things plotly and Dash allow me do!
However, I am facing a weird issue when using a colorscale (viridis) on a scattergl plot with around 20k data points. The vast majority of points gets colored perfectly fine, but every now and then I stumble across a seemingly miscolored dot, usually a “too high” color. My colorscale has values from 0.0 to around 4.5. When I look at the click data for the point it shows the correct value for marker.color (e.g. 0.49), but the color is somewhere near the max. I included a picture, with an example data point. The coloscale is mapped to the value named “outlier ratio”.
Initially I thought opacity is the problem (which frankly might have been part of it), but I am still facing the issue at opacity = 1. Another thought was a mixup between multiple points with the same coordinates, but the points with the issue have unique coordinates within my dataset.
So far I couldn’t come up with a reasonable explanation why this is happening. I also could not recreate the behaviour in a simpler dataset, so I don’t really know how to go on with this. Maybe anyone has some advice on how to figure this out? 
