Geoprofiling

[mpriestnall]

JTR could have lived in Essex and used the rail network to commute in and out of the East End.


Just throwing that grenade into the pot



RD

And why not? He could have lived anywhere within reasonable traveling distance of the murder sites.

[JeffHamm]

How useful actually is Geoprofiling? Do serial killers always kill in the area where they live? Or do they as some have suggested tend to kill slightly away from where they live? For example a neighbouring area.

It's not quite that simple. First, despite how it is presented in the press (and often by researchers, sadly), it's not just about where the offender lives. While the offender's residence is often a very strong anchor point, that's not always the case anymore than thinking that the spouse is always the offender just because they commonly are! Spatial analysis is like making a list of locations similar to how one lists people to investigate (spouse/ex-spouse, family members, friends, etc) in an ever widening social circle (so that list is like organising "social space" into zones, while spatial analysis is about organizing physical space into zone). For some offenders, the main anchor point of their offending can be their place of work, for others it's an pub or area of town they frequent regularly, and so forth. In those cases in particular it's easy to see how their not living in the hot spot isn't a failure of the analysis, but rather a failure to understand what the analysis actually indicates.

The purpose is to suggest a search pattern, it's not telling you what evidence you will find. It is useful to the extent that the suggested search pattern locates the offender sooner than if you searched randomly. The idea, of course, is that because the locations are not chosen at random by the offender, but rather the offence locations are related to something about how offenders choose locations (generally the idea is that they choose areas they are familiar with, and much of that familiarity comes from their day-to-day lives, which in turn indicates they will be familiar with areas around their "anchor points"), then if you applied the routines to randomly chosen locations, the routines would do poorly. But if you applied the routines to actual offenders, the routines would do better than searching random locations.

That's one of the things I looked at in the paper I wrote on this topic. I generated random series of crime locations, placed the offender's residence randomly, and used a couple of different routines to see how well they locate the offender. The routines define a search area and divide that into 40 zones (each 2.5% of the total search area). It is possible that the offender's location falls outside the defined search area (in the random data that occurred about 19% of the time). I then plot the percentage of random offenders that are located as you follow the search pattern, so how many are located by only searching zone 1? how many if you search zones 1&2? and so forth.

I then looked at real offenses; I had the data from serial arsonists in New Zealand, but having looked at other types of crimes, like serial murder and serial rape, the results don't change. And the real offenders are far far more likely to be located than randomly generated series (82% were located by zone 20, which is 1/2 the search area). In fact, over 25% of real offenders were located in zone 1 (an area only 2.5% of the total search area), and 50% were located by zone 4 (10% of the total search area).

One thing, though, is that for the real offenders the only anchor point I had was their residence. Some of the arson locations were areas around beaches, or lakes, etc, and I would not at all be surprised if the offenders were known to frequent those locations for normal activities, like swimming, surfing, camping, fishing, hunting, etc. So the anchor point for their offenses wasn't their residence, but rather a location connected to their hobby. What that means, is that the figure below underestimates the performance.





So, do some offenders follow the "commuter pattern" (meaning, travel well away from the area of their day-to-day activities to offend)? Yes, of course. Israel Keyes is probably the best example I can think of as representing a commuter - someone who travelled specifically to offend (he was caught when he committed his last offence in Alaska where he lived). Long haul truckers are people with a residence that moves with them. However, those examples are far far far fewer than the "marauder" pattern, which is one who offends in the area of their daily activities. And for those, which are the vast majority of offenders by far, the spatial routines do pick up on information about the offender's spatial habits.

- Jeff

[JeffHamm]

Hmmm, thinking a bit more about the commuter pattern, I can think of only one case where someone commuted to the same location to meet victims ( Colin Ireland). He lived about 40 km east of London, but would travel into the city to a particular gay bar to find victims. He claimed to be heterosexual, and I am not familiar enough with the details of his life to know if that is considered to be true, but if not, then that pub may have been where he went to be far enough away from where he lived so as not to be seen living his secret life. In that case, the pub becomes part of his non-criminal activities, making it an anchor point of the marauder type. However, if he truly only went there as part of his criminal activities, then it is truly a commuter pattern. And if that is the case, he is the only example of a commuter I can currently think of where the crimes are relatively clustered and not spread over hundreds of miles. Interestingly, the spatial analysis does tend to hotspot the area where the club was located (top 2 or 3 zones I think). It is still a sizable search area, but nowhere near as big as the area that would encircle all his crimes.

So, given the rarity of that pattern (clustered crime locations but still a commuter), it is far more probable that JtR lives fairly close by; not necessarily in the hot spot, as I think a pub is the more likely anchor point in this case (could be wring of course), but even if so, I would think he lived within 5 minutes walk of the area.

- Jeff

[Fiver]

The issue with a commuter is that he would be limited to the operating hours of trains. These murders, excluding Liz's, would be after trains had stopped running. Now, if he stays over in the bolt hole then leaves the next morning, maybe.

That is a significant issue if the killer was not based in London. It's still an issue if the killer was not based in the East End - there would be far fewer cabs at that time of day and the killer would greatly increase his risks by giving the driver a chance to notice bloodstains, etc.

There is a lingering issue that may relate to where he lived and what he did......the gap between Annie then Liz, and from Liz to Mary. Polly and Annie were killed within 2 weeks. The that annoying gap until LIz, who is cut only once, and Kate, who is the only "Ripper" victim on City land. What Im hinting at is this...what if Liz and Kate were killed by different people and neither was Jack? Shall we then assume he doesnt kill in the East End from early September until early Nov? Or should we question whether Mary was in fact a Ripper victim....which could suggest that after Annie, that man just left the area?

If Israel Schwartz was telling the truth, then Stride probably wasn't a Ripper victim. If Schwartz was lying or mistaken, then it's a coin flip over whether it's an unrelated killing or if the Ripper was interrupted. Use of a single cut is a trivial variation in signature.

Killing Eddowes in the City of London police jurisdiction doesn't even imply that it was another killer. Period police did not agree how many Ripper victims there were, but they agreed that Eddowes and Kelly were Ripper victims. Modern experts in signature analysis agree - Eddowes and Kelly were Ripper victims.

If the Ripper moved after killing Chapman, then why didn't he continue killing in a new location?

[The Rookie Detective]

That is a significant issue if the killer was not based in London. It's still an issue if the killer was not based in the East End - there would be far fewer cabs at that time of day and the killer would greatly increase his risks by giving the driver a chance to notice bloodstains, etc.



If Israel Schwartz was telling the truth, then Stride probably wasn't a Ripper victim. If Schwartz was lying or mistaken, then it's a coin flip over whether it's an unrelated killing or if the Ripper was interrupted. Use of a single cut is a trivial variation in signature.

Killing Eddowes in the City of London police jurisdiction doesn't even imply that it was another killer. Period police did not agree how many Ripper victims there were, but they agreed that Eddowes and Kelly were Ripper victims. Modern experts in signature analysis agree - Eddowes and Kelly were Ripper victims.

If the Ripper moved after killing Chapman, then why didn't he continue killing in a new location?

Excellent post Fiver.

I think you make some very important points that are integral to focussing on the identity of the Ripper.

Modern criminal psychologists and experts will likely say that a serial killer will typically begin killing nearer to home and then move further afield as they become more confident.
But what if with the Ripper; the opposite was infact the case?

Confidence didnt seem to be an issue for the Ripper.

The gap between Chapman and the double event is key because there is a definite shift in focus and geography....but rather than moving further away from home; was the Ripper instead moving closer to home?
With the double event he moves out of Spitapfields and then kills in St Georges in the East and then the City of London on the same night.
The double event locations tell us so much more about the Ripper than is perhaps realised.

When we combine that with the deliberate shift in attention and focus from the press and the forming of the WVC, it perhaps forms another intriguing coincidence.


In terms of the Geoprofiling; the focus has been on areas of Spitalfields, particularly around Flower and Dean St and Dorset Street, but with the inclusion of the double event and the killer choosing to essentially kill 2 women outside of Whitechapel, he hints at his locality being more expansive and transient.

But the key for me is that the killings of Eddowes and Kelly felt more...personal.

Eddowes claimed somewhat flippantly (allegedly) to have known the identity of the Killer and MJK was butchered inside her own room.
Eddowes also gives her name as Mary Kelly...a slightly odd coincidence in itself.


Did the murders of Eddowes and then Kelly bring the killer closer to home rather than further away? (as criminal psychologists would state)

Eddowes and Kelly felt personal and a lot closer to home for the Ripper....does that then run parallel to the Geoprofilling?

Nichols and Chapman forming the Eastern and Northern boundaries to his kill zone respectively...

And then with the double event followed by MJK, the killer comes closer to his home, both physically and mentally?


Lots to ponder


RD

[Lewis C]

Excellent post Fiver.

I think you make some very important points that are integral to focussing on the identity of the Ripper.

Modern criminal psychologists and experts will likely say that a serial killer will typically begin killing nearer to home and then move further afield as they become more confident.
But what if with the Ripper; the opposite was infact the case?

Confidence didnt seem to be an issue for the Ripper.

The gap between Chapman and the double event is key because there is a definite shift in focus and geography....but rather than moving further away from home; was the Ripper instead moving closer to home?
With the double event he moves out of Spitapfields and then kills in St Georges in the East and then the City of London on the same night.
The double event locations tell us so much more about the Ripper than is perhaps realised.

When we combine that with the deliberate shift in attention and focus from the press and the forming of the WVC, it perhaps forms another intriguing coincidence.


In terms of the Geoprofiling; the focus has been on areas of Spitalfields, particularly around Flower and Dean St and Dorset Street, but with the inclusion of the double event and the killer choosing to essentially kill 2 women outside of Whitechapel, he hints at his locality being more expansive and transient.

But the key for me is that the killings of Eddowes and Kelly felt more...personal.

Eddowes claimed somewhat flippantly (allegedly) to have known the identity of the Killer and MJK was butchered inside her own room.
Eddowes also gives her name as Mary Kelly...a slightly odd coincidence in itself.


Did the murders of Eddowes and then Kelly bring the killer closer to home rather than further away? (as criminal psychologists would state)

Eddowes and Kelly felt personal and a lot closer to home for the Ripper....does that then run parallel to the Geoprofilling?

Nichols and Chapman forming the Eastern and Northern boundaries to his kill zone respectively...

And then with the double event followed by MJK, the killer comes closer to his home, both physically and mentally?

[bolding added]

Hi RD,

There's as much of a shift in geography from the Nichols murder to the Chapman murder as from the Chapman murder to the Stride murder. And if he moved closer to his physical home with the double event, then he moved back away from it with the Kelly murder.

Geoprofiling takes the locations of the double event murders into account, except that there may be geoprofiles that have been made to cover the scenario in which Stride isn't a Ripper murder.

[Lewis C]

That is a significant issue if the killer was not based in London. It's still an issue if the killer was not based in the East End - there would be far fewer cabs at that time of day and the killer would greatly increase his risks by giving the driver a chance to notice bloodstains, etc.

Hi Fiver,

Michael's point that if The ripper was a train commuter, he would have had to wait for the first train the next morning raises the question, when would the first train the next morning have been? I suspect that at least in Chapman's case, he wouldn't have had to wait very long, and maybe not in Nichols' and Chapman's cases either.

[New Waterloo]

It is suggested that possibly JTR lived a little more South of the murder locations.

A little while ago I did a bit of research on the name P Johnson mentioned in SB registers and have suggested an identity of this man in a thread as below
P Johnson identified (Possibly)

02-28-2024, 04:19 PM
A while ago Trevor (I think it was him) came up with the name P Johnson 'Said to be Jack the Ripper' in Special Branch ledgers.

Not a lot to go on by I have located a Peter Johnson who looks interesting for a couple of reasons;

1881 Census

210 Cable Street, St George in the East

Peter JOHNSON b. 1842 Seaman born Germany

Elizabeth JOHNSON b. 1842 born Whitechapel

Peter JOHNSON b. 1878 (son 3 years of age)

Some interesting observations;

My research seems to indicate that Elizabeth's maiden name was ROWE (wasn't there a ROWE connection to Barnett)

but more interesting is living at the same address is a family with the surname MACKNESS

William MACKNESS b. 1825 Holme Huntingdonshire

and other family members

I dont really know what all this means but this Johnson is a Seaman and Abberline was married to a Mackness (first marriage) from that area.

ALSO

In 1891 JOHNSON is at another address in Cable Street and I cannot locate the rest of the family

I am not great at research but there do seem to be some possible connections here.

Importantly he lives in Cable Street which I believe is a bit further South (stand to be corrected)

I would really appreciate any thoughts on this character

Thanks

NW

[Lewis C]

This is probably nothing more than an interesting coincidence, but George Chapman's barbershop in 1888 was at 126 Cable Street, which was also listed as his residence in a 1889 London directory. So based on the numbers, that sounds like it would be less than a block from where Johnson lived in 1881.

[JeffHamm]

Interesting. Yes, Cable Street is just south of the east/west running tracks below Berner Street. Not sure of the numbering though.

One thing, though, someone living on Cable Street would be making some hard to explain choices. If we include Stride as a Ripper victim, then JtR has headed away from Cable Street in order to get to Mitre Square and the Eddowes murder. On the other hand, if Stride is not a Ripper victim, it removes the one crime location that is near Cable Street. The other things that make me pause for thought, is that Chapman's murder, which I think occurred somewhere around 5:25am type thing (not universally agreed upon, and I'm not wanting to start that debate here, just explaining the context in what follows), is one of the furthest locations from Cable street, meaning he's got the longest distance to traverse at the most dangerous times (busy morning activity, light, etc) of being spotted (and potentially showing blood). Also, the Kelly murder, while more likely to have occurred when it was dark, is another time when he might have the most signs of blood on him, and is again one of the most distant locations from Cable Street. While this isn't a "sure thing", it does strike me that for both of those murders (the Chapman one in particular), that he would only progress to murder if he felt comfortable enough about getting out of sight (home) quickly, minimizing his at risk exposure time. That makes me think that JtR had somewhere he could go that was nearer to the Chapman and Kelly crime scenes than the others. Also, it might suggest why he goes from Berner Street to Mitre Square - he's heading in that direction. It might even be the distance from home that resulted in him aborting things in the Stride case (again, these last ideas presume Stride is indeed a JtR victim).

Certainly the above ideas are highly debatable, and I'm not putting them forth as any sort of "this is how it was" argument, rather just following a few ideas that grow if we make certain assumptions as our starting points. Make other assumptions, you'll get somewhere else, otherwise known as an alternative hypothesis.

Hmm, for the Georeferenced Historical Maps, is there an overlay that shows the street numbering? I've not found one, but I'm sure there must be one that shows this information, even if only for buildings at intersections. Ideally one that has decent detail. I'm not too concerned if the street names have been changed. Even if the detail isn't great I can always transfer the information. I know there's the thread with the link to the detailed JtR map, which is great, but there are times like this when it would be interesting to locate new locations that get mentioned. I have a working version map where I tried to map out various locations with suspects, etc, but it's gotten a bit messy and had a few errors on it, such as I had marked Chapman's residence on Whitechapel road, but he lived on Cable Street at the time of the murders. I would like to re-do that in a clean version, on a decent looking overlay, as I use it to plot the spatial analysis outputs and it can be interesting to see which, if any, suspects are in locations of interest. As I recall, though, Cable Street isn't in a zone of interest, but as I've said many times, this type of analysis isn't evidence, and while it "works" more often than not, there are times when the not wins. Also, I tend to think the hotspots being highlighted are reflecting a non-residence type of anchor point.

- Jeff

[New Waterloo]

Another slight coincidence (or maybe not) is that Louis Diemschutz stabled his horse and cart at George Yard, Cable Street. Presumably he was dropping off his goods at home before taking his horse to the stable yard. So there is a firm connection with Cable Street and Dutfields Yard and the club. I am not sure where George Yard is on Cable Street would be interesting to know. A numbered map would be most handy. Can anyone assist with this please. Thank you.

NW