Data: The blood of the smart city

The idea of ‘Moore’s law’ caught everyone’s imagination as we saw computers increase in power as they decreased in size and cost. The actual underlying observation was that the number of transistors on integrated circuit boards doubled every two years. Interestingly, but coincidentally, we are seeing a similar thing happening with data. In December 2017, 54% of the world’s population (4.15 billion) were Internet users (1).

Internet usage is never benign.

The Internet has evolved into a personal data transfer system. We have online accounts for virtually every online interaction. Accounts, like banking, ecommerce, gambling, and so on, each containing personal as well as behavioural data. We also each own, on average, 3.6 connected devices (2). There is so much data that we have had to build new technologies to create it and analyse it. Big data is one of the phenomenons of the modern age.

Data, however, is not an abstract thing. Data is a representation of us, as individual humans, or as a group of humans. It presents a technological front that shows how we behave, how our families behave, our friends, our enemies. Data used in a smart city context is our reflection of our human face and our behaviour.

Big data is the lifeblood of the components that make up our smart cities. It is being hailed as the “new oil”, “like gold”, and the saving grace of future technologies. However, it can also be a poison chalice. Data is only as good as the value it offers and this value is determined by a number of things including in-built bias, the trustworthiness, and the value of these data. It is also an area where governance needs to build a strong guard. Data privacy is often an afterthought. It seems like we have to sacrifice privacy for progress. This should not be the case and does not have to be the case. Strong control and governance of smart city data will bring rewards in trust and respect by those citizens that data represents.

IoT and cyber-physical systems: The beating heart of the smart city

The Internet of Things (IoT) or connected devices, is one of the manufacturing success stories of the modern era. These devices are used in both a consumer and commercial basis. This dual usage model means that IoT devices are part of the infrastructure of the smart city. IoT devices act like the beating heart of the city, pumping the blood (data) around. But, IoT devices can also generate data. For example, the health apps market is expected to grow at an astonishing CAGR of over 44% to 2025 (3). This is an important area of technology that can help in the optimization of continuously stretched healthcare services. As our population grows in size, it also grows in healthcare needs. Non-communicable diseases are as harmful to the smart city as communicable ones were to our first cities. Diabetes 2, for example, is predicted to grow by 165% to 2050 in North America (4). Healthcare apps, which communicate health data to healthcare workers, can help to alleviate the stress on first-line and hospital care.

Cyber-physical systems (CPS) are needed to create an ubiquitous, always connected, service infrastructure within the smart city. Where the cyber (integrated network, e.g. Internet) and physical (e.g. an electrical grid) meet -data feeding the collaboration of the two. There are a number of technical challenges in creating a homogeneous layer in the smart city and the cyber-physical is one of them. As we discover more ways to connect the cyber and the physical we need to be cognizant of their thirst for our data to analyse and use to optimize the service. Much of the smart city will be dependent on a variety of CPS, including smart transport, smart grids, smart medical services.The risk of data breach is a serious consideration in the design and implementation of a CPS. A CPS is a critical infrastructure in its own right. Some may also interact to become co-dependent. You can see where this is going. Touch one CPS and you may well expose others. An example of how this is impacting our lives currently, in a non-smart scenario, is in the world of credential stuffing.

Credential stuffing example: A recent Federal Trade Commission (FTC) investigation into data exposure affecting a accountancy firm, TaxSlayer, set a new security and compliance precedent (5). This involved the use of ‘Credential stuffing’ practices which involves indirect credential exposure. For example, company Y has user credentials compromised. This results in company Z suffering a data exposure using those compromised credentials. The FTC ruled that company Z was as liable as company Y. This has created a rule of interoperability that will need to be understood and accounted for in co-dependent CPS within a smart city context.

Again, respect for privacy of the data of an individual has to become second nature to the smart city design — compliance at least will dictate this.

Artificial Intelligence and Machine Learning: The brain of the smart city

All of these data will be used, somehow in the smart city because information provides insights into patterns and trends which can be used to optimise operations and facilities. If you have enough of it, aka big data, you can get a pretty accurate picture of whatever it is you are exploring. So, for example, smart grids, with enough information at hand, can be used to determine peaks and troughs in electricity need and to then adjust output. Machine Learning (ML) is a subset of Artificial Intelligence (AI). ML takes the data generated by the health apps, or smart meters, or Internet-enabled cars, etc., and uses these data to spot patterns and learn how to optimise the given service. For example, NVIDIA have developed the smart video which handles big data analytics and applies machine learning to video streams. They have partnered with 50 AI city partners to utilize the technology to improve areas such as smart transport. There are expected to be 1 billion of these intelligent cameras by 2020 (6). That’s an awful lot of data generated, analysed, and acted upon. The system will replace human interpretation, replacing it with machine learning algorithms — with an expected improvement in accuracy and speed. This city brain will process a lot of our personal data, including visual data about our movements.

Machine Learning requires data to spot patterns and trends. The analysis of big data gives city services the information needed to be highly responsive to the needs of its citizens. It also uses these data in services to build more optimized responses to service use, helping to enhance the experience and improve sustainability. One area that is being explored as suitable for AI and machine learning is in the personalization of services. This requires that personal data is collected and aggregated, before being used as a profiling tool. ML tools that personalize experiences are already in use in marketing, for example. Here they are used to tailor online sites, displaying products that users are expected to like from their predicted profile.

In a smart city, the same type of algorithms can be used for other purposes. For example, a study by three UK universities looked at the application of various ML algorithms to cycling and weather as a means of creating personalized services within a smart city (7). The study was based on the collection, aggregation, and analysis of big data. The study concluded that a “combination of ML, IoT, and Big Data offer great potential to developers of smart city technologies and services.” This study was done without the need for data that could directly identify an individual. That is not to say that with effort, correlated data, perhaps with GPS from mobile devices, could be used to re-identify individuals. Also, it is not too big a leap to imagine that even more tailored personalization, or more accurate results, could be obtained by using directly identifiable information.

One of the other concerns about machine learning and AI is the possibility of default bias built into the very algorithms that are supposed to improve accuracy. If the training set itself is skewed towards a specific expected outcome, then the result will be skewed in turn — in fact, the resulting bias may well be amplified. There have been several studies in this area including, “Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints” (8) which looked at how training sets contained gender bias which then became amplified when used in an AI system.

The use of bias in AI might also amplify privacy concerns. An example of where this type of bias and control has crept in, was the use of Microsoft’s ‘Tay’ Chatbot which was trained using real-world tweets. The problem arose when people started tweeting racist and misogynist comments to Tay who then played back those sentiments. Although this may seem extreme, it only takes one rogue programmer to build-in bias. Similarly, privacy issues could arise from biased training sets. Privacy is more than the exposure of personal data. It is the exposure of our very being — our beliefs, our views, our political leanings, and so on.

Privacy in the smart city is much more than revealing your name…

My next article takes these ideas further: Smart City Factors 2