Your algorithm that reports duplicate reports is wrong. It assumes that the LAST located report is the real thing and prior nearby reports are duplicates. This is incorrect. You should assume that reports are NOT duplicates at all. Let me give an exact case. On the highway n12/n3 intersection on the corner of woodmead (suburb) in Johannesburg RSA, starting at -26.047675193550557, 28.093193441959258 and ending at -26.03960699164605, 28.067755371505946, on the date 31 Nov 2021 at around 4pm, there were THREE stopped vehicles at the roadside. Your system recorded them as THREE REPORTS at the LAST location, rather than three separate reports. In actual fact, there were THREE stopped vehicles. The first one was roughly at -26.047675193550557, 28.093193441959258 and the last one at -26.03960699164605, 28.067755371505946. The middle one in between was IN THE ROAD at 120kmh. It was NOT visible around the curve. HIGHLY dangerous. I tried to report it but it was assimilated into the last vehicle which was about ½ km FURTHER. I suggest you assume that reports are duplicates within a 50-100m difference max. My reporting accuracy is VERY accurate, I report precisely where the incident is. We have BAD drivers, and LOTS of accidents and incidents. DO NOT assume it is a duplicate.