Safer and smarter: The power of synthetic data in smart cities


Modern society runs on data. It is everywhere. From booking a restaurant reservation to planning your commute, so much of our daily functioning is reliant on information that allows the technology around us to run successfully.

The unprecedented availability of data – coupled with new abilities for how it can be utilised – drives constant innovation and research. This ability to deliver more advanced data-driven insights is massively accelerating growth and this is being seen now more than ever within smart cities.

The overall value of the global smart cities market is expected to exceed $2.7 trillion by 2027, with the official artificial intelligence market in smart cities predicted to reach $298 billion. Facilitating this forecasted growth is dependent on data-driven insights and real-world data alone is not up to the task. Huge amounts of data are needed to supply information and communications technology across a wide range of services, including healthcare, transport and energy.

It is evident that data and the ability to process it quickly is vital to seamless functioning of the connected, intelligent networks underlying cities. This demand is only going to continue to grow as the number of IoT devices worldwide is predicted to almost triple to 29 billion by 2030. The current principal objective is therefore the ability to supply accurate data that addresses corner cases for safe and intelligent systems at the rate required.

Safety and sustainability

Digitally transforming cities is driven by a desire to make them more liveable and the key components of this are sustainable, smart and inclusive technologies. It is essential to place economic growth and sustainability parallel to one another to not only build prosperous cities but also to address environmental challenges that are a threat to quality of life.

Safety is of the utmost priority and city officials aim to provide citizens with improved living conditions. Real-time video monitoring and sensors for object detection can help to reduce crime through crowdsourcing of data about crime, active tracking and allowing authorities to respond more rapidly. While real-world data alone is well-suited to deal with easily predictable, typical scenarios, in the case of random acts of violence or crime, there is a clear gap for dealing with anomalous situations. Implementing technology that is trained to capture and recognise such incidents can allow for human supervision to be diverted elsewhere, where it is more urgently required.

Smart technology can be deployed to assist in a number of sustainable functions, including maximising energy efficiency, monitoring air pollution levels and providing data on how to eliminate waste. In urban environments, there is arguably an added need to be more mindful of environmental impacts. Smart lighting for example, can be adapted for energy-efficiency, only being utilised when human presence is detected and also, can be trained to account for edge cases such as severe weather conditions in an attempt to improve security but also reduce energy consumption and expenditure.

It is indisputable that data will play an increasingly pivotal role in safety and sustainability in the discussion on smart cities. Data and AI have the ability to transform how safety protocol is conducted and this is reliant on the technology being flexible. The data used to power these systems is key to achieving this. Ultimately, providing a holistic understanding of the population and deepening trust between citizens and city officials through leveraging data is a requisite to safer and smarter cities.

Why timing is key

All of this technology is reliant on data insights and predictive insights. Global examples of how smart cities are utilising technology are ever growing but with this comes the potential for bottlenecks that can slow down operations and be accompanied by scalability challenges. While real data is useful, it can be time-consuming and labour-intensive to accurately collate at the speed necessary if we wish to ensure safer and smarter cities evolve at the rate demanded. This risks impeding the development of smart cities and is where synthetic data comes in.

Synthetic data is generated from computer systems with the intention of resembling real data both statistically and structurally. It enables us to decipher large volumes of data from a multitude of sources quickly and accurately. Synthetic data not only allows for computer vision systems to be trained more rapidly but also cost-effectively and ensures confidentiality. As more data is artificially generated for any edge case needed, the potential for creating more refined data sets that better resemble real world use cases gets better.

Real-world data still has a role to play and there is no taking away from its integral contribution to training AI systems. A combination of the two data sources is fundamental to innovation. Real-world data’s foundational aspects combined with synthetic data’s potential for scaling delivers robust AI models covering the key use cases while being thoroughly tested against edge cases. Therefore, it’s vital that we popularise jointly training with synthetic data to develop smarter and safer cities on a global scale.