I have several shapefiles with data all around the USA. All of them are well under the 50k point limit. But clustering will only work on small sections of each file. AGOL mysteriously decides that it only wants to cluster, for example, the bottom third of the points in the South. And on a different shapefile, perhaps the points in the Northeast, and that's it. And nothing I do changes its mind. Something in AGOL is deciding which points from my files it's going to cluster and then it's saving that information, because even if I open the same map on another computer, the SAME group of points clusters, and it ignores all the other points. Zooming in does not help. When clustering is turned on the missing data never displays unless I'm zooming in one of the areas it wants to cluster, or I turn off clustering. I've been forced to go back to displaying every single point in order to get all the data to display.
I know it CAN work because about 10% of the time, it DOES work! The rest of the time it's ignoring two thirds of my data. I know clustering is processed in the browser, but things like clearing the cache, trying different computers, and trying different browsers has no effect: the same small group of points are the only ones that cluster. Am I missing something? Is it just still buggy?
Here's a location that happens to be on the border of the Clustering vs Nonclustering zones. This is in Santa Fe and every location in the USA north of the invisible line on the left map is not clustering. The points just disappear:
![](/legacyfs/online/399293_clustering.JPG)