COPLINK CompStat Analyzer automates crime data analysis

     Since its inception in New York City in 1994, CompStat — short for comparative or computer statistics — has gained considerable recognition for its role in knowledge-based law enforcement. Used by police departments across the country, CompStat enables police and residents alike to see where crime is happening and respond accordingly.

     Yet, according to Robert Fund, product manager at Knowledge Computing Corp. in Tucson, Arizona, and a retired 30-year veteran of the Tucson Police Department, traditional CompStat solutions are costly. That's because they have been internally developed and are thus highly customized. As a result, data analysis is a manual process which demands multiple staff members and can take anywhere from days to months.

     Enter Knowledge Computing's COPLINK CompStat Analyzer, a solution that automates data analysis processes from compilation and comparison to visualization and real-time analysis. This capability has made the new product key to the recently rolled out Southern California Gang Emergency Operations Center (GEOC), an initiative that will pull together resources from many disparate criminal justice and community groups in the region.

     CompStat Analyzer uses numbers from an agency's existing databases, which underlie the COPLINK technology. "COPLINK solutions, including CompStat Analyzer, use any underlying data source an agency elects to map and migrate for search capability," Fund explains. "Typically an agency will map an RMS (records management system), JMS (jail management system, including mugs), citations, auto accidents and the like, which all refresh based on a pre-determined refresh time mechanism — anywhere from once a week to every few seconds."

     That enables query results to be returned in mere seconds — and to be parsed into geographic and other visual display forms. Moreover, the data collection phase itself is automated and ongoing. So, says, Fund, "Each time a user asks to analyze a given period of time, he is presented with data from the most current to the most historical."

     These capabilities will be critical to the GEOC, a knowledge-based policing project that will connect law enforcement and justice groups with local government, social services, educational institutions, and community and faith-based organizations. (Los Angeles County Sheriff Lee Baca, mutual aid coordinator for emergencies in Los Angeles and Orange counties, modeled the GEOC on disaster response.)

     Not only will the GEOC have access to what Los Angeles County spokesperson Lt. Chris Cahhal calls "traditional information" — RMS, CAD (computer aided dispatch) and citations — it also will be able to access parole and probation, Section 8 housing and other data. "In the near future, analysts and deputies countywide will have access to a number of non-traditional databases," Cahhal says. "These databases include information on demographics, education, economics, employment and housing."

     Why such a broad spectrum? "In evaluating a problem, it is important to get all the facts prior to determining a course of action," Cahhal explains. "For example, should we find that crime has a nexus to probationers, then it would be best to work with the probation department in enforcement actions." The additional database resources will also help its human services side "get the big picture, formulate a game plan and deploy resources based on the facts presented."

     COPLINK users had asked the company to develop a tool that would help them improve community-oriented policing effectiveness. To that end, says Fund, "CompStat allows an agency to determine where and when the most need for increased police presence exists." That means that unlike the other COPLINK solutions, CompStat Analyzer is not tactical. "It's designed to give the user a strategic view of what has happened historically within their operational environment," says Fund. "CompStat looks at aggregate data — the big picture — to see overall trends."

     Users can filter queries by time, location, crime type, time of day, day of week or month, gender, race, object (such as vehicle type), division, beat, activity, officer role, or any combination thereof. Such information can then be imported into a number of analytic tools, including graphs, charts and GIS-based maps — and analyzed in any way, including spatially and temporally. "For example, what is the city's most frequently stolen car?" says Fund. "Where and when are the most auto thefts occurring? How many auto thefts occurred in June 2008 compared to June 2007? How do auto thefts in one area compare to auto thefts in another?"

     This makes it easier for police to decide how to deploy resources. "For example, a university police department might use CompStat to map activity and determine a bike theft 'hot spot,' "explains Fund, "and in turn, dispatch additional personnel to a specific location at a specific time — the place and time the most bike thefts occur."

     Another important aspect of CompStat Analyzer is its ability to keep track of performance not only on the departmental level, but also on shift, unit, or even beat and individual levels. "CompStat creates 'institutional knowledge,' " says Fund. "In other words, a patrol officer can determine what types of crimes occur most on his beat or where a 'hot spot' would require extra patrol — knowledge that would normally take years of experience to acquire."

     COPLINK products have always been modular. Its base solution is Detect; and the others, including CompStat Analyzer, run on top of that in a plug-and-play format, each independent of the other.

     While CompStat Analyzer is planned for use in the Los Angeles County's Crime Assessment Center (CAC) — an all-crimes fusion center that provides what Cahhal calls the "analytical support for the GEOC" — it is expected to also be employed in every sheriff's department station; and if technically possible, in every radio car, too. "In the past, units have relied on a manual process to pull up statistics and create pin maps," Cahhal explains. "With the automation of the CompStat module, shift managers and even deputies can view crime statistics to do their own analyses."

     Fund says the modules are so intuitive that officers need hardly any training to use them. In fact, users can "play" with the software as much as they want because the data remains at its source — the underlying database. Product demonstrations often take several hours, not because the system is hard to use, but because officers are so interested in what it can do.

     Because CompStat Analyzer is a stand-alone module, non-law enforcement community members can see and understand accurate, up-to-date aggregate data without access to incident details. "[This] provides leadership with a dynamic tool for public presentations on crime trends and performance while significantly reducing preparation time," says Fund. "CompStat enables agencies to provide the public with information regarding crime trends in the community."

     "Since the GEOC concept is not limited to law enforcement, the CompStat module will be used during presentations to the human services side of the GEOC," Cahhal points out. "The graphic display of crime trends will assist in conveying the extent of the crime problem to our non-law enforcement partners."

     A strategic partnership between Knowledge Computing Corp. and ESRI now enables COPLINK modules to integrate with ArcGISServer 9.2 and 9.3. "COPLINK can directly leverage an agency's GIS investment and provide advanced mapping for criminal investigations," says Fund. "This will allow greater scalability for large clients and support more map data and image formats."

     The enhanced functionality extends to information sharing with certain external systems. "The COPLINK program has interfaces to the federal data sources OneDOJ and ICEPIC," Fund explains. "These are two-way interfaces, allowing queries from both sides. In addition, COPLINK can export and absorb information in the Global Justice XML Data Model (GJXDM), which is NIEM [National Information Exchange Model] compliant and allows our users to push data to N-DEx."

     Cahhal says the GEOC concept is part of a larger regional data sharing initiative that began (along with Los Angeles County's COPLINK deployment) in December 2007. In Los Angeles County, the sheriff's department's COPLINK node joins two others: the Los Angeles Police Department node, and the Regional Terrorist Integrated Information System (RTIIS), a node for the remaining municipal policing agencies in Los Angeles County that is expected to be fully functional by the end of 2008. "In addition to these nodes," Cahhal says, "The Los Angeles Sheriff's Department is sharing data with Orange County's COPLINK node (ILJAOC), ICE, N-DEx, and is close to signing a data sharing agreement with San Diego County (ARJIS)."

     COPLINK CompStat Analyzer does not need to be customized to an agency, so it remains comparatively affordable. Yet despite being proprietary, it can work with any RMS built on any database. In agencies with more complex requirements, such as the Los Angeles Sheriff's Department, it does require a degree of customization. "L.A. County has 140 million records in its node," says Cahhal, who says this caused some initial performance problems with CompStat Analyzer. Now that a database upgrade is complete, however, beta testing has resumed with the module's latest version.

     For its actual operation, says Fund, "The COPLINK application requires no more bandwidth than it requires to view the typical CNN.com homepage. Depending on your activity within the application, the bandwidth can range anywhere from about 10K for basic result screens to 200K for complex Incident Analyzer and CompStat displays."

     Many police departments may not have the infrastructure to support the use of COPLINK modules like CompStat Analyzer in patrol cars. Fund says more agencies are installing Wi-Fi systems that will enable this kind of usage. In the meantime, COPLINK Mobile exists for departments that want their officers to access data via limited-bandwidth devices, such as mobile phones and PDAs.

     As DeLorenzi, Shane & Amendole wrote in "The CompStat Process: Managing Performance on the Pathway to Leadership" in 2006: "It is undeniable that the core management theories of CompStat, 'directing and controlling,' have been demonstrated to be effective means for controlling crime. But the CompStat process also has an inherent opportunity for developing leaders and improving the leadership process ... when used effectively for accountability and problem solving, [it] can be a means for developing potential leaders and promoting cooperative and creative leadership." CompStat Analyzer's easy access to records and its automated processes make this even more of a reality for agencies that use it.

     Christa Miller (www.christamiller.com) is a freelance writer based in southern Maine.