Big Data Opportunities to Transform Cities
Municipalities of all sizes are destined for big changes from the big data phenomenon
American poet A. R. Ammons once said, “A word too much repeated falls out of being.” And although the term big data sometimes seems to be too often repeated, it’s not about to fall “out of being” when it comes to the transformation of today’s cities. In fact, it won’t be long until many who live in cities large and small look back and recall the ways in which the use of big data and big data opportunities made their cities smarter, more efficient, and more capable than ever.
Big data opens up many possibilities for cities. From enhanced patient outcomes in local hospitals to tax collection, road safety, crowd control, traffic flow, garbage collection, and opinion polling, the list goes on. Big data offers municipalities opportunities to transform policies, ideas, and outcomes like never before.
To share the art of the possible, advanced big data technologies such as Apache Hadoop are changing the economics of purging and retaining data that can result in municipal governments. They have unprecedented opportunities—and face extraordinary challenges—to proactively change the lives of their citizenry, all through the use of data.
Painting pictures of opportunity
There are many aspects of daily life in municipalities that can be significantly impacted by the opportunities big data offers. Though there are quite a few, some key areas include the health, safety, movement or traffic flow, and revenue systems that sustain inhabitants.
Health: Getting ahead of the next outbreak
Consider how municipalities typically handle an outbreak of influenza or food poisoning. They react. Infected people start to arrive at hospital emergency rooms with a crescendo of symptoms that, in retrospect, confirm that a suspected illness has become an outbreak. Local medical officers then respond with after-the-fact notifications, forecasting activities, and summaries of net impact.
In a big data world, this entire process gets inverted; forecasting becomes nowcasting. For example, in a big data analytics-led municipality, medical officers use an extraction protocol that understands verbiage associated with illness to interrogate social forums in an effort to identify people who are expressing sentiments related to how they are feeling. Public expressions of an upset stomach, vomiting, fever, chills, and so on are all clues. Frequent comments about food from a certain vendor and associated symptoms are clues to a potential food poisoning outbreak. Does this interrogation activity raise a privacy concern? Sure, but the people are putting the data out there for all to see.
Sentiment analysis can be tricky. Think of the word “orange.” If people living in upstate New York are basketball fans, they’re likely to think of the Syracuse Orangemen. Anyone with a cell phone in Europe will probably have a well-known telecommunications provider in mind. In Ireland, it has religious meanings; in Florida, the word is an icon for the entire state. In many other places it’s just a color.
Beyond words with multiple meaning, contextual analysis is key. For example, the medical officer in the previous example is more concerned with expressions of vomiting after eating food from a specific vendor or restaurant location than if those expressions relate to last night’s Metallica concert.
Using this kind of approach, municipalities are going to be able to identify a flu outbreak through social media well before their citizens show up in hospital emergency rooms. A real-time dashboard that classifies both syntax and grammar extractions from various data sources can reveal growing evidence of a potential problem and help healthcare practitioners proactively determine where precious resources should be allocated (see Figure 1). Such classification systems work on data as it is born into the world, and they can be referred to as in-the-moment analytics systems.
Figure 1. In-the-moment classification for a disease-tracking dashboard
Additional insight can be gained and profiles can be built to model pending situations. For example, profiles of people currently expressing flu symptoms in Jackson, Mississippi, can be used to identify their interests in sports and in their professions self-categorized as IT (see Figure 2). In this way, imagine trying to trace an outbreak of food illness and being able to understand the connections to a national IT conference with an evening sporting event where the hot dogs turned out to be not so hot.
Figure 2. Profiles from gained insight to model specific outcomes
These heuristics were reverse-engineered from the social profiles of the people expressing the onset of some sort of illness. They allow classifications to be created that trace the source of the illness by understanding characteristics as they relate to the afflicted—hobbies, jobs, and so on.
Safety: Nabbing big data numbers
Police departments are transforming safety initiatives and using big data, not only to catch the bad guys committing crimes, but also to implement protocols that help prevent crime. For example, the Boston Police Department leveraged Twitter to communicate emergency procedures and to clarify inaccurate press reports during the hunt for the marathon bombing suspect. The department saw Twitter as the new instant press that millions of people were using to follow events—the bombing and ensuing chase to hunt down the suspect—both textually and visually. Twitter is a big data tool that has been used to make heroes and ruin careers within a framework of only 140 characters.
Moreover, the world self-organized—in real time—around the bombing event through the use of the #bostonmarathon hashtag and later the #bostonstrong hashtag. Today, the Boston Police Department is using big data technologies and systematically correlating and analyzing all kinds of data, from parole records to video footage to local event data—all to keep its city safe.
The Rochester Police Department in Minnesota uses big data as a cornerstone of its intelligence-based policing approach. For example, data associated with a vehicle being driven by a person of interest can be used to connect the dots between anything or anyone that has ever been linked to that specific vehicle. The information can be delivered through on-the-spot analytics to officers within 15 seconds. The goal is to surface nonobvious relationships and associations that are attached to a vehicle of interest.
The basis of this technology comes from the gambling tables of Las Vegas, where casino management wants to establish a risk score of its employees as it relates to fraud. For example, was a blackjack dealer’s sister’s roommate once involved in slip-and-fall scams in the area? The obvious question is, “Does this dealer have a record?” The difficulty delta between the obvious questions and questions that seek to establish relationships is akin to the difference between finding a needle in a haystack and finding a needle in a stack of needles. The latter metaphor indicates a potential benefit of big data.
Movement: Going from A to B and sometimes C
With the ever-expanding population of urban centers and increasing traffic congestion, a city’s ability to move its population is a multifaceted problem that has measurable economic and quality-of-life impact on a city. Big data opens the door to public transportation awareness solutions that help improve on-time performance and provide real-time bus arrival information to riders. By continuously analyzing bus location data to understand traffic conditions and predict arrivals, cities can deliver transportation systems that move people—both physically and from a quality-of-service perspective.
For example, the city of Dublin in Ireland has implemented its Intelligent Transportation System that is designed to provide speed and traffic-flow measurements, travel time estimates, and statistical aggregations of current traffic indicators as they happen. This solution provides Dubliners with real-time visualization of the arrival times for 1,000 buses on many routes and stops.* This information has enabled Dublin to optimize bus routes and stop locations, not only making its transportation system more efficient and trustworthy, but also significantly driving up its ridership and revenues.
Singapore offers another example. Ever been there and tried to get a taxi during a heavy rainfall? It’s almost impossible. Why? The answer may be surprising. A big data study that focused on this problem used Global Positioning System (GPS) devices in taxis to gather location awareness information. Combining geo-positional information with weather data revealed that taxis in Singapore simply stop moving during heavy downpours. When the big data net was cast further, the reason became clear. Because taxi drivers in Singapore are responsible for their own deductible payments, they chose to simply pull over when driving was deemed too risky from a financial perspective. The confluence of position data, weather data, fare accruals, and insurance policies revealed these fascinating insights.
Revenue: Levying taxes to support infrastructure
Taxes. For many it can be a dirty word, but tax collection is the financial lifeblood of a city. Without them, how can municipal initiatives be carried out, roads maintained, schools run, or parklands protected? Big data helps increase the ability to quickly spot anomalies, enabling municipal collections to reduce the tax burden gap—the difference between what taxpayers owe and what they voluntarily pay.
Having more data that yields greater insights than previously imagined can profoundly affect the motivation of anyone who considers filing fraudulent tax returns—for example, a house assessment that increases because of a finished basement that should be voluntarily disclosed. In this case, a filed building permit application for the finishing of a basement would be a key piece of information to determine if an act of assessment fraud had been committed. The possibility that these documents are shared today is highly doubtful. In addition, accurate and up-to-date property assessments can be collected through an analysis of traffic patterns, crime information, and other socioeconomic factors that uncover points of view not previously visible.
Taking the next path
The potential benefits of big data for municipalities are well beyond the scope of a single article. From safety, to revenue collection tools, to transportation, to voting logistics, to opinion gathering, and everything in between, the big data phenomenon is destined to be a shift, lift, rift, or cliff. The big question is, which of these will it be for any one city? Which path does a city take, and can big data get the city there faster than without big data—even when it’s raining?
Please share any thoughts or questions in the comments.
* “Dublin City Council,” IBM case study, August 2013.
|[followbutton username='IBMdatamag' count='false' lang='en' theme='light']|