What Are These?

The Murder Accountability Project has developed an algorithm capable of detecting serial killers who target multiple victims using similar methods of killing within a specific geographic area. This technique can be useful to police in detecting difficult-to-see patterns over a period of several years or longer.

The algorithm is based upon a reasonable premise -- an active serial killer can reduce the normal (or expected) clearance rate for groups of similar victims killed through similar methods. The algorithm organizes more than 700,000 homicides into about 100,000 clusters by generating a unique Murder Group number based upon the geography (either county or metropolitan area), victims' gender and method of killing.

The algorithm looks for clusters with very low solution rates.

By moving the "% Solved" slider, you can adjust the sensitivity of the search. Higher settings will produce more suspicious clusters, but also increase the odds that other factors influenced the failure rate of police investigations. By concentrating on a particular city, careful manipulation of the "solved" selector demonstrate what kind of homicides are the most difficult to solve.

To see the victims of "Green River Killer" Gary Ridgway, set the cluster maps to the years 1980 to 2000, and the "% Solved" selector to an unusually low rate of homicide clearance, perhaps 25 percent or 33 percent. You will see a large circle above the Seattle area. These are women who were killed by "other or type unknown" weapons. Since most of Ridgway's 48 female victims were found out-of-doors, medical examiners usually had difficulty determining the precise cause of death.

The Murder Accountability Project believes these clusters with lower-than-expected clearance rates have an elevated probability of containing serial killings. But they are not proof of the presence or absence of multiple-victim offenders.

Rarely are all of the victims within a cluster the handiwork of serial killers. Police investigation – including physical evidence, offender confession, or witness testimony – is the best evidence that a cluster of homicides may be linked.

To see the algorithm written in SPSS syntax (which can be translated into other statistical software systems) click here.

These visualizations were developed by Haneesh Marella and donated to the Murder Accountability Project. Contact Haneesh through his LinkedIn account here.