Reclaiming independence in spatial‐clustering datasets: A series of data‐driven spatial weights matrices