Читать книгу GIS Research Methods - Steven J. Steinberg - Страница 28

A headline on a local news website reads, “Westside Mugger Caught!” Given that you work on the Westside, you feel a great sense of relief as you begin to read the article. The past few weeks have seen a rash of muggings; every couple of days, another victim was attacked, and it seemed as though the assailant was a step ahead of the police. You have always wondered how the police catch up with criminals, and as you read the story, you come across a sentence that piques your interest: “We never would have caught the person behind these attacks without our new CompStat system,” stated the chief of police. The article goes on to explain that CompStat is a computer-based analysis system built around crime statistics mapped in GIS.

Interesting. You begin to wonder exactly what the journalist means by mapping crime statistics. How would that help catch a criminal? You have always found maps to be interesting, and they certainly help you find your way when traveling. You have even heard about those maps to movie stars’ homes you can buy in Hollywood, but you don’t recall ever seeing a map to criminals’ homes (figure 1.1).

It turns out that the GIS behind CompStat wasn’t exactly used to find the home of the criminal, but almost. The police took advantage of a variety of basic information, or data, about the area in which the crimes were occurring (figure 1.2), along with information about the locations of each of the muggings as they were reported. As the locations of the crimes were mapped, some interesting patterns began to develop.

Figure 1.1 An example of a publicly available crime map from the Hollywood area of Los Angeles, California. This web-mapping site integrates crime data from police departments around the country and is powered by the Esri ArcGIS for Server. Courtesy of the Omega Group, San Diego, CA. Basemap data from Esri, HERE, DeLorme, IPC, METI/NASA, USGS, EPA.

Figure 1.2 The dashboard view provides detailed, specific crime information to police chiefs and precinct commanders. Courtesy of the Omega Group, San Diego, CA.

For example, all of the muggings occurred within two blocks of an ATM machine. That seems sensible to you; the mugger might well have been targeting people who were getting cash. The attacks were always late in the evening, after ten o’clock, and the victims were always confronted on streets that had little traffic. What streets don’t have lots of traffic at that time of night? Perhaps residential areas, where folks are in bed? Maybe. But wouldn’t someone hear the commotion? A more likely area is around the financial district, where everything closes at five in the evening and there’s not a lot going on at night.

You start to realize that by looking at some basic map information, you might be able to narrow down areas that meet a particular profile that seems to be developing. But as you think about it, you wonder, “Aren’t there lots of areas on the Westside where there’s little activity in business districts at that time of night? And ATMs? It seems there’s one on almost every corner. What else could have helped the police get the bad guy?” It turns out that the crimes were clustered in a ten-block area. Perhaps the mugger lives near that area, or, better still, he probably lives near the middle of that area, so he didn’t have to walk too far to find his victims.

Of course, the police knew other things that helped them narrow down the suspect. They knew who in that area had a record for mugging, robbery, burglary, or other similar crimes. They knew if any recent parolees lived nearby. They may have had other clues that matched the modi operandi of known offenders. Odds are this wasn’t someone who woke up one day and decided to become a mugger—someone like this probably has a history.

As you ponder all this, you begin to understand how a system like CompStat would be so helpful. Of course, if you could somehow put all of these data together on a map, defining areas that meet the given criteria, you might be able to narrow the search area down to something manageable. Sure, you might not come up with the criminal’s home address, but you would certainly know where to put extra police on the beat to catch him. But one thing still bothers you: the complexity of getting all of this information onto a map and doing the analysis to get to this point. Wouldn’t that be a major task? It was hard enough for you to draw a readable map for your friends to find their way to your new apartment for your last Super Bowl party! It’s much easier to direct your friends to one of the many online mapping tools to find your house.

It must be that computer thing the police chief mentioned in the article, that geographic information system, that performs such a complex task. It all sounds very complicated. But you are intrigued and want to find out more about these geographic information systems. Maybe they could be useful in other ways. After all, if you can use them to narrow down locations of criminals, what other kinds of analysis might they be useful for?