Policing Better Data
Tell us about your organization.
Humberside Police is one of 43 police departments in England and Wales, and is responsible for policing a sizable county in northeast England. All told, our department consists of approximately 4,000 police officers and related personnel.
What problems were you facing?
We had a number of challenges, many of which a typical organization or business wouldn't encounter. For example, while a customer will rarely provide a business with fake contact information, a [criminal] suspect does this all the time. This presents some unique challenges. Additionally, we collect data from numerous outside sources, including other local and U.K.-based law-enforcement agencies, such as Scotland Yard. So in the end, we had this heterogeneous collection of suspect and criminal data that was coming in and being entered into disparate systems. We had a central database to record criminal events as they occurred, but it was separate from many of the other databases. To make a long story short, it was hampering our ability to provide our officers with timely and accurate information, reducing the efficiency of our intelligence unit, and hindering our ability to operate with other law-enforcement agencies.
Why did you select Informatica?
In the autumn of 2005, we realized we needed a tool to cleanse and match the data coming in from the disparate systems into a central data warehouse. We did a proof-of-concept and tested several data-cleansing products. We were shopping for a product that would show us where the information was [questionable], and help us pick out areas that needed to be corrected -- so this was very much a data-governance issue. In the end, we were most impressed with Informatica from both price and functionality standpoints.
How did the implementation proceed?
We ran into numerous technical glitches associated with a data-quality project, such as determining data distribution and integration woes, but all in all, it went off like clockwork. Our internal databases were built using Microsoft SQL Server, with an Autonomy search engine running on top of it. We use Informatica to cleanse information as it's coming into the data warehouse, info such as names, contact information, addresses, records -- all that good stuff. Then, the Autonomy search engine distributes that cleansed info out to our police officers.
To ensure end-user buy-in, we involved the officers and our intelligence unit from the very beginning and gave them demonstrations on how the new system would work. We also worked closely with our intelligence unit to determine which data is leveraged the most and to map data-governance and -stewardship issues.
What have been the main rewards?
Since implementing Informatica, the Humberside Police department has seen the accuracy of its vehicle-registration numbers climb from 83 percent to 98 percent, and that's just one form of data that's seen a big improvement in accuracy. Our intelligence unit has been able to analyze data like never before, and it's given us the ability to verify case leads with law-enforcement agencies throughout the country.
There are hundreds of specific examples. Oftentimes, information is collected under duress or pressure, and that can lead to inaccurate data. We're now using Informatica to cleanse that information and match it up with complete records to give our officers the most up-to-date information.
5 FAST FACTS
AGE OF THE INITIATIVE? 1 1⁄2 years
WHO WAS INVOLVED? Myself, our IT managers, and the department head of our intelligence unit
BEST IDEA? Coming up with a separate data force to oversee all of the data governance issues
BIGGEST SURPRISE? The inaccuracy of the data when we started the implementation
BIGGEST CRM MISTAKE MADE? Failing to understand the impact dirty data can have on an IT implementation