Paving the Way with Analytics

Turning information overload into actionable intelligence

• By Kevin Taylor
• Sep 01, 2019

How many of you remember learning about the Great Chicago Fire of 1871? Some of you might even recall the ditty that sprung from the legend of its origin: Late last night when we were all in bed Mrs. O’Leary hung a lantern in the shed And when the cow kicked it over She winked her eye and said, “It’s going to be a hot time in the old town tonight. Fire! Fire! Fire!”

While there’s no irrefutable proof that the cow was the culprit in the devastation that left miles of the Windy City in ruins, imagine how different the outcome might have been if Mrs. O’Leary had been connected to an electric grid to light her property or if she had a monitored system that detected the situation and communicated out to emergency responders.

Technology—from street lights to telephones—has been at the heart of modern cities for more than a century. Even the idea of “Smart Cities” isn’t new. There is a publication on the concept dating back more than a quarter century. (The Technopolis Phenomenon: Smart Cities, Fast Systems, Global Networks, D.V. Gibson, G. Kozmetsky, and R.W. Smilor, eds., Lanham, MD: Rowman & Littlefield Publishers, 1992). Though actual implementation of the concept is far more recent, cities have been using advanced technology to address the challenges of densely populated urban centers for years.

Thanks to the Internet of Things (IoT), there are now rapidly emerging platforms that leverage interconnected devices to help cities oversee everything from power plants and water supply networks to mass transit, emergency communications, and law enforcement.

The unfortunate downside to all this interconnectivity is that amassing these large, and seldom uniform, data sets can lead to organizational paralysis rather than the desired results of knowledge, action, and measurable positive outcomes.

This leads to the suggestion that what we really need are “Smarter Cities,” ones in which we can intelligently sift through that vast amount of IoT-generated data, quickly extract the most essential information, and act decisively based on what we’ve learned. And this is precisely where analytics excel. Analytics can be a powerful force multiplier when it comes to efficiently processing volumes of multisourced data—whether thousands of video cameras, environmental sensors, access control switches, or other devices.

While there are countless analytics on the market today, I’d like to focus on those related to video surveillance. These analytics generally fall into three categories: detection, data collection, or forensic review.

Proactive Detection

Most video surveillance systems are used primarily in post-event investigation. When coupled with analytics, however, you can transform them into proactive detection systems that alert operators in real-time that something of interest is taking place. Some of the more common examples of detection analytics include intelligent motion detection, crowd gathering, loitering, and object left behind.

There is tremendous upside to the early warning these analytics provide—whether it’s notifying police that a crowd is building in an unanticipated area so that they can respond to a possible fight before it escalates into an all-out riot; or letting event organizers and public safety officers know that a temporary barrier at a parade or festival has been breached so they can work together to mitigate a potential threat and re-establish the barrier; or even recognizing that a suspicious object has been left behind near a cultural landmark or mass transit depot so that an operator can quickly validate the alarm and initiate a pre-planned emergency response.

Most detection analytics operate based on a set of filters that the user must initially configure for the specific use case. With the emergence of Deep Learning, though, some analytics can be trained over time to improve their accuracy.

However, no detection analytic can be 100 percent accurate. Their performance and delivered value largely depend on how you define the specific use case and the desired outcomes. Working through this pre-deployment with all the stakeholders will ultimately lead to the best selection and provisioning of hardware and software to accomplish your city’s goals for detection. You’ll not only miss fewer critical events but also avoid excessive nuisance alarms.

Analytics can also play an important role in city planning. Because cities are constantly evolving, municipal officials are continually monitoring, measuring, and evaluating conditions so development decisions can be made based on actual current data.

For example, pushing a video stream through an analytic application can help city departments collect specific data on pedestrian, vehicle, and bicycle traffic. The collected data then provides a basis for informed decisions about crosswalk and bike lane placement—even mass transit schedules.

Another emerging trend is to bridge data, such as lane-by-lane vehicle counts and travel times, in real-time to traffic signal control equipment. There are now cities that use analytics to minimize traffic jams by moving traffic signals off hard-set timers and operating them dynamically based on real-time conditions captured by intelligent video cameras.

Expeditious Forensic Review

As more police departments partner with the private sector to enhance video coverage of public spaces, analytics are becoming essential tools for intelligently processing the massive amount of video data flowing into command centers from multiple sources. In the past, it would take officers days, and even sometimes weeks, to sift through all the video related to an incident before finding the few critical seconds of footage.

Now, instead of searching for a needle in a haystack, there are analytics that enable officers to query the data set using specific identifying parameters such as someone wearing a green shirt or driving a red car in a certain window of time.

Relevant video pops up on the video monitor significantly expediting the investigation.

Analytics can also help law enforcement recognize patterns over time, like the migration of crime from one neighborhood to another. Some police departments are even using analytics like dwell time and vehicle/pedestrian traffic patterns to pinpoint the location of drug dealers.

Given that many police departments are understaffed and light on resources, it becomes even more important to include analytics in their crime-fighting arsenal. Swifter, more efficient forensic review frees officers to spend more time building relationships with the community and fostering trust.

The Burgeoning Portfolio of Analytic Solutions

As more metropolitan areas embrace the concept of Smart City, we’re bound to see a rise in analytics development and performance. New companies will emerge, eager to capitalize on the opportunity to promote their suite of products. But before a city considers the burgeoning portfolio of analytics, officials need to ask themselves:

• What pain point are we trying to resolve or what goal are we trying to achieve? • Is analytics the best way to address the issue or can existing tools do the job? • Can the analytics be easily scaled with equity for the betterment of the entire community • Is the solution provider in the relationship for the long haul with a strong record of service and support?

Analytics hold great allure for cities exploring ways to improve efficiencies and create a safer, more secure environment for their citizens and visitors. And with constant development in code and algorithms, today’s analytics are continuously becoming more reliable and accurate in detection, data collection, and forensic review. In addition, they’re becoming more efficient in terms of processing requirements. When you couple these achievements with more robust processing power of edge IoT devices, video analytics for cities becomes even more appealing.

These developments broaden the selection of analytic applications that can run directly on edge IoT devices and sensors, overcoming the challenges of solution architecture and cost that are common when application processing must be de-coupled from the sensor (run on different hardware such as a hardened PC or in the cloud).

will likely see more analytic offerings in the near future that will run at the edge and scale well for cities—a true 1:1 ratio architecture of sensor and application processing in a single device.

Going from Smart to Smarter

With more people flocking to urban centers than ever before, cities will continue to look for innovative technology to help offset the strain on resources and improve the quality of life for their citizens. By adding analytics to their operational toolbox, municipalities can raise the bar and transform their smart cities into even smarter (and safer) ones.

This article originally appeared in the September 2019 issue of Security Today.