Fra Carnevale is one of the most mysterious painters of the Renaissance. He has only nine works attributed to him, in which architecture is the protagonist. His best-known work is “The Ideal City” (circa 1480), where he depicts a metaphysical urban landscape, a utopia of order, and geometric purity. At a similar time, an Italian architect nicknamed Il Filarete dreamed of the Sforzinda, another utopian city. Indeed, the very term “utopia” comes from the title of Thomas More’s work on an ideal society and architecture. Fortunately, five centuries later, imagination has a technological ally to make things easier: big data applied to mobility and urban planning.
The design of cities and the movement of pedestrians and vehicles are intimately linked. So much so that the type of mobility, from horses and carriages to electric scooters, has shaped cities.
With both factors considered, mobility can be managed in the medium and long term through urban planning or modulated in the daily activity of the population, either with the frequency of public transport, the distribution of shared vehicle fleets, or information on traffic jams and accidents. Thus, in this article, we will talk about
Big data and mobility in urban planning: the example of CityScience Lab
The future, if we are to follow population indicators and UN forecasts, is eminently urban. And within this exponential growth of infrastructures, big data is destined to be a vital tool of the smart city.
An analysis using big data allows us to understand the traffic load of different areas and routes and the peak demand for public transport. For example, so-called heat maps are generated in this way, showing locations at risk of congestion.
However, big data makes it possible to go even further: the wealth of information available offers the possibility of changing street signs or even rationalizing the growth of a city and its connections.
Much of this planning is simplified through the visualization of big data, as in the CityScience Lab project at the University of Hamburg.
The initiative, which the MIT Media Lab has supported, brings together open data from the Hamburg city council on building typologies and surface areas. It also incorporates data from traffic infrastructure and air quality or anonymized data from smart homes.
The software makes it possible to generate what-if scenarios like the impact of a population increase and the distribution of residents. Along these lines, CityScience Lab was used to accommodate refugees from the Middle East in 2016.
Within the project, the use of a 3D model of this German city also stands out. The concept, known as a “digital twin,” consists of replicating all the city data in real-time. In this way, any change in traffic, power lines, signage, or the construction of new urban roads is reflected in its digital version.
One of the first cities to debut a digital twin was Singapore, with a project explained here. Other significant examples include Wellington, in New Zealand, or Shanghai, the latter with more than 2,300 virtual km2.
Big data and daily urban mobility
Urban planning tends to be a long-term strategy, but mobility management is also linked to the daily heartbeat of a city. Arriving at the workplace can be an adventure in this fast-paced ecosystem, and finding parking can quickly become a nightmare. In addition to the time spent, poor mobility means more air and noise pollution. And here, too, big data has something to say.
A primary application of big data in urban mobility is the distribution of parking spaces. Some studies indicate that up to 30 % of traffic jams are related to vehicles looking for parking.
The systems currently being developed work with IoT devices. In this way, data from parking meters and vehicles entering or leaving particular zones can be combined to establish occupancy levels.
However, the use of big data in mobility also makes it possible to implement predictive systems. In this way, it is possible to offer suggestions about the times and areas where there is a greater probability of finding a free space without tracking each individual space.
Prediction of accidents and traffic jams
As we have seen in the case of parking, big data is an excellent resource for assessing urban mobility and anticipating events. One of its latest applications has come out of another MIT lab and consists, fundamentally, in predicting the risk of traffic accidents.
Based on deep learning models, the system is the result of a collaboration between MIT’s CSAIL and the Qatar Artificial Intelligence Center. The researchers combined historical accident data, satellite information, road maps, and GPS information as a database. In total, about 7500 km2 of urban areas and 7.6 million km of GPS routes were processed.
One of the model’s advantages is that it is applicable in areas without previous accident records to pinpoint trouble spots. For now, it is an ongoing project, but it could be used to support urban planning.
Traffic jams are another area where predictive models based on big data can contribute to improving mobility. The Swinburne University of Technology in Australia has worked closely with the company Intelematics to create a new model capable of predicting traffic jams with an accuracy of up to 96% and anticipation of up to sixty minutes.
Car-sharing and big data
If predicting traffic accidents or anticipating available parking spaces are applications still under development, the use of big data in the case of shared vehicles is at a much more advanced stage.
Scooters, motorcycles, and shared cars – primarily electric – represent a paradigm shift. However, they also offer numerous logistical challenges.
Addressing issues such as vehicle recharging, breakdowns, and availability depends mainly on reliable and up-to-date data. In parallel, knowing (and predicting) user behavior patterns requires obtaining and processing data with big data models applied to mobility.
One of the examples in this regard is ACCIONA’s work in the field of shared electric scooters. The company manages the fleet of bikes and anticipates peaks and areas of demand using big data. Its moto sharing app also calculates the amount of CO2 saved on each trip and the possibilities of combining a motorcycle trip with public transport or even with other nearby competitors.
The future of urban mobility
There is a jocular saying about nuclear fusion because of its protracted development: “It is the energy of the future and will always remain so.” Something like that could be said of autonomous cars, which look set to remain the future of transportation for some time to come. With any luck, however, they will not meet the fate of flying cars.
The only sure thing is that the use of big data in this field will be vital, as it is estimated that autonomous cars will generate up to 40 terabytes of data per hour. As explained in Venture Beat, rather than advanced radar or LiDAR-like sensors, the key will lie in deep learning systems based on a simple video signal. Today, Tesla is processing data from millions of vehicles on the road to develop its driverless models.
In short, transportation and urban mobility are destined to undergo the same revolution as other sectors such as e-commerce or social networks with the advent of the big data era.
And the application of this technology promises to translate into greater sustainability and efficiency of the means of transport we move around with, be they scooters, motorcycles, cars, or even bipedal engineering systems.