tl;dr a short biography of Thomas Hargrove, leader of the Murder Accountability Project (MAP). He is a former crime reporter who has often involved statistical analysis of crime databases in his work. MAP provides geographic profiling, paired with other similarities to highlight potential serial killers.

The article begins with Hargrove's biography, expands into some previous attempts, and concludes with a contrast of another member of MAP's board, who began a similar effort with the FBI in 1983. Throughout, this article relays some aspects and categories among serial killers and their victims.

This reminds me of a recent discussion [2] about another statistical source: real-time health data. Some alleged the data sinks aver on the grounds that such streams could engender legal liability. One of the interviewees alludes to this, but false positives are presumably more expensive in this case. Which sort of sucks. I don't envy police, and certainly not whichever county is dead last in unsolved murders, as that's apt to be a poorer community making the same sorts of compromises we all need to make but on a more emotionally charged scale [3].

[1] http://www.murderdata.org/p/how-to.html [2] https://news.ycombinator.com/item?id=15717136 [3] https://www.youtube.com/watch?v=F9juReoJxI0?t=29m35s

Interesting to read about an organization working on a hard problem, but that hasn't yet solved it. I wish it gave more evidence about his 'algorithm' that showed the patterns. It didn't sound like his data background was advanced, so I'd hate to think that he's using an underpowered tool to solve an important problem. I keep wondering if someone with a solid understanding on data science could speed up his work considerably.

Interesting because the University of Chicago's Crime Lab does work like this to track the epidemic and virality of murder as it happens across Chicago. Worth a listen on an amazing podcast I highly recommend called Hidden Brain: https://www.npr.org/2017/02/21/512035150/on-the-knifes-edge-...