I am trying to create strongly related clusters of events based in two dimensions…the date/time of event and it’s location based on UK Easting Northing.
Ideally I want to find out the group of events that are in a similar time frame, either all around the same date point or sesame day and time window as each other (Fridays , around 20:00), and in a similar area. However I do not want to include those that are not “strongly” part of a cluster.
The analogy I can think of would be delivery spots, where I want to look at the 20,000 spots I already have recorded and see where to hot spots are but not have the one or two outside the man clusters influencing the group’s to be reviewed.
Any good starting point on the? I have started with k means but could not understand how to creat only the “strong” clusters.