How Artificial Intelligence (AI) Will Help You Understand Your Data

Big data and AI are highly complementary to each other – big data sets enable AI to leverage its strengths in machine learning; and through its processes, AI renders complex big data understandable by discovering relationships that would have gone unseen even by today’s big data tools.
Winbourne Consulting (AI) Artificial Intelligence
Governments as a whole, and public safety agencies in particular, have significantly more data available to them today than just a few years ago. Indeed, it is an almost unfathomable amount of data compared to what was available 25 years ago when paper forms ruled.  As “big” as this big data is, it is about to get even bigger!  A number of current trends and emerging technologies indicate yet another unprecedented growth in data.  Some of the components of this surge in data include:
Emerging Technologies
  • Body-worn cameras
  • Social media
  • Drones
  • Fifth-generation (5G) cellular networks
  • Smart cameras
  • Internet of Things (IoT)
  • Smart Cities 
  • Next-generation 9-1-1 (NG911)
  • Public Safety Broadband Network/FirstNet
Ideally this amount of data would lead to improved, data-driven decision-making through the normal data -> Information -> insight -> decision-making process where raw data, properly collated and analyzed, is transformed into information which leads to insights on the business process that leads to enhanced decisions driving service improvements.  This process is the very foundation of intelligence gathering initiatives such as CompStat and similar programs in law enforcement and other areas of government.
Historically, the majority of data received by decision-making bodies has been previously defined in one way or another. As such, this data fits nicely into traditional database structures which render it more amenable to analysis and description of relationships between various types of data.  As the volume of this data has grown, new tools have been developed that can handle the volume while delivering acceptable results.  There are numerous big data tools available today that do an excellent job of analyzing large data sets that traditional tools do not have the ability to effectively manage.
However, as both the amount of data and types of data continue to grow even those tools could be taxed.  The growth in unstructured data – data without a format, such as text or video – in particular presents a further challenge.  As we deploy next-generation 9-1-1 across the US, images, videos and voice records will become part of the 9-1-1 call for service and dispatching.  As more advanced technology such as body worn cameras and videos from drones or crime scenes are deployed, those files will also become part of the public safety information domain. The resulting sets of structured and unstructured data available will present a challenge to any individual or analyst trying to discern meaning and driving service improvements. 
Given this increased level of complexity, public safety will need a new approach to intelligence analysis to properly leverage these data streams that are extremely large and diverse in their makeup.  Artificial Intelligence (AI) presents a potential way forward to dealing with these vast amounts of data.  Indeed, as an article in the Sloan Management Review put it :
“The convergence of big data with AI has emerged as the single most important development that is shaping the future of how firms drive business value from their data and analytics capabilities.” 
Big data and AI are highly complementary to each other – big data sets enable AI to leverage its strengths in machine learning; and through its processes, AI renders complex big data understandable by discovering relationships that would have gone unseen even by today’s big data tools. 
For example, one of the areas where AI is making great progress is in the field of video analysis.  As the amounts of both private and public video cameras have expanded, more and more of this material is being made available to public safety, often in real-time mode.   While it is fairly simple to gather the recordings from a small set of cameras around an incident for review, it is quite another challenge for a human to monitor every camera feed in real-time to react to incidents as they occur.   While the former activity can be very useful in solving a crime, the latter could conceivably stop a crime in progress.   
Of course, it is impossible to monitor numerous video feeds in real-time to look for indications of criminal or unusual activity – unless you’re an AI tool.  Through the use of machine learning and neural networks, an AI-enabled system learns what normal crowd behavior looks like at a particular location and time – for example, a train station looks different at rush hour than at 2PM.  With this “knowledge”, the system can then distinguish between the normal behavior of a crowd versus one where something unusual is occurring, and alert the proper authorities.  
A related, though more controversial application, is running live facial recognition (LFR) on crowds.  In LFR the software is assessing video streams in real time and comparing faces in the crowd against a database of known individuals.  LFR has been extensively deployed in some locations in China, but has also been used to some extent in other countries, including in the US during national security events, such as the Super Bowl.   The Metropolitan Police in London, i.e., “Scotland Yard”, recently deployed mobile-mounted cameras running LFR in high-traffic areas , albeit not without controversy.
What’s Your Strategy?
We are clearly on the verge of a new era of leveraging AI and big data to gain better business intelligence in public safety and enhancing the ability of public safety organizations to increase their effectiveness.   There are numerous potential applications of AI and big data depending on the data sets that are, or will be soon, available to any public safety agency, and the use of that data will vary from jurisdiction to jurisdiction.  In order to make the best use of this data, most experts recommend that an organization develop a strategy that delineates its goals and objectives based on the types of data available.  Once a strategy is adopted a phased implementation period can begin.  Depending on the amount and types of data to be analyzed this phase can last from weeks to months before a program is underway in earnest.  Of course, once underway, the program can be enhanced with additional data sources or analytical tools as they become available.   
Agencies need to start this process today to ensure that they make the right investments in this emerging technology that will enhance their level of service delivery while avoiding dead-ends.  The Winbourne Big Data team has the experience and expertise to support you in this effort.  To begin the conversation, contact us at

Recent Posts

Sign up for our Newsletters

Scroll to Top