StreetCred: Big Data Finds Fugitives

By

Traffic citations are critical for funding police departments, courts, and municipalities. Unpaid citations or missed court appearances result in warrants and additional fees, but finding or arresting fugitives and collecting unpaid fines is difficult as well as time and manpower intensive. As a result, the process often stops there. In Texas alone, there is almost $1 billion in outstanding warrants, with $300 million in Houston and $100 million in Dallas.

StreetCred Software is a Big Data solution designed by two Texas police officers, one a technology industry veteran who later joined the force to cut down on that backlog by collecting and analyzing all available information to optimize and manage the entire warrant lifecycle. When tested over a year within a Texas police department, StreetCred helped warrant officers increase their successful contacts by more than threefold while decreasing costs per contact by nearly three quarters.

A good misdemeanor warrant officer will spend days planning routes, printing maps and photos, checking addresses, doing paperwork, and other research and preparation before he or she can actually go out and make contacts and arrests. As a result, in smaller agencies, the warrant officer will typically only spend one to two days per week in the field, during which he or she can expect to make five to seven good contacts where they speak with or confirm the address of a. Aside from a lack of basic route planning, this rate is hampered by false, misleading, or simply incorrect information as well as attempts to avoid law enforcement by fugitives. To determine which warrants are worthwhile and which are misleading, or which home or work address to check, a veteran warrant officer will consider dozens of small factors such as, available information on the fugitive’s employment, the given neighborhood, and the type of violation or crime committed. This undocumented, manual, and repetitive process helps officers determine which warrants to pursue and how. Failures in this complex and subtle process mean missed contacts and missed opportunities.

StreetCred Software uses Big Data analysis techniques to automate these procedures and hunches. The software aggregates court, police, and law enforcement database records together with open source and other information. It alerts officers of potentially dangerous fugitives by correlating misdemeanor data with felony and violent history data from sources traditionally unavailable to warrant officers.

It compiles those records, storing them on server in the agency’s data room, and accessed by a secure connection from the standard laptop in the cruiser, so that he or she doesn’t need to assemble and carry around several boxes of papers, and analyzes that data, automating the conclusions and hunches of a good warrant officer.

StreetCred can do that because it’s built for cops by cops, and its c0-creator, David Henderson, has run a warrant department for a Texas agency, pursuing misdemeanor and felony fugitives. In his 15 years in law enforcement, David has tracked and arrested over 600 violent felony criminals including a child murderer featured on the television show America’s Most Wanted, and has worked that experience, as well as the insight of his colleagues, into StreetCred’s algorithms.

Using all of the available information, StreetCred scores the thousands or tens of thousands of warrants that it works with on a scale of 1 to 100, with the highest scores being the ones most likely to result in an easy contact or arrest. This helps the agency prioritize their warrants and not waste time with warrants that are no longer relevant, have insufficient or false information, or are otherwise unlikely to yield results. Such warrants may also be important and would then be pursued by the police department, possibly with the help of StreetCred, but are not a good choice for quickly clearing the backlog or collecting revenue. StreetCred also plots warrants on a map with their scores and whether the address is home or work to help with route optimization. Once contact has been made, StreetCred uses the information it gathers to help fill out the paperwork associated with the contact, freeing up more time. As a result, when StreetCred was implemented with a Texas law enforcement agency, it took the average number of contacts per day from 5 to 22 during the first year.

StreetCred’s interface also simplifies making contacts. All of the relevant data on a fugitive can easily be accessed through the touch-screen laptop in the officer’s cruiser, which means that it is intuitive and can be used quickly on a touch screen. Relevant information is prominently displayed in such a way that officers can reference it while driving without getting distracted, information important for officer safety such as the picture of the fugitive stays on screen, and all the buttons are large for easy access, with no information more than one click away. Agencies can also add their own information, such as overlaying crime hot spots or sex offenders on to the fugitive map.

StreetCred has been tested for a year in a small city near Dallas where it achieved impressive results. It reduced the cost-per-contact by 68% by lowering required research time, printing costs, and vehicle operations costs, and more than doubled the clearance rate per hour worked of misdemeanor warrants.

Fugitive hunting software can be worrying due to both privacy and information security concerns, but StreetCred Software takes both into account. The algorithms never factor in race or gender when scoring warrants, and the software helps keep track of contacts and arrests to prevent abuses. StreetCred is also less concerned with finding good warrants than throwing out bad ones. Numerous features are built in to ensure that sensitive information isn’t leaked such as extensive physical and information security measures in the infrastructure, making sure that no information leaves StreetCred without encryption, and special views on the mobile versions of the software to prevent private information from being seen by onlookers when an officer is walking around or asking neighbors if they’ve seen the person in their picture.

As with all innovative solutions, however, StreetCred raises several questions and concerns that will need to be addressed with time and wider deployment. The most obvious is that some of its initial success may simply be a result of plucking the lowest hanging fruit, the warrants just waiting to be collected. Once officers begin to work through the backlog, the more challenging warrants will remain, which may lower the success rate. Still, given current backlogs, this would be a good problem to have. Another concern is the focus on misdemeanor warrants. While this approach makes sense given tight budgets, some of these features would be valuable for felony warrants as well and could do serious good there.

StreetCred is a great example of how Big Data can be leveraged for public safety. As with predictive policing, StreetCred helps officers do more and better with less time and money, which is critical now with shrinking departments and budgets. While StreetCred already handles an otherwise overwhelming amount of information, it’s working to make the data even bigger by enabling national cooperation and information sharing among law enforcement agencies. As StreetCred, which is currently deployed at a handful of agencies, continues its launch, we can look forward to more creative and necessary applications of Big Data for law enforcement, government, and public safety.


Original post

Leave a Comment

Leave a comment

Leave a Reply