The Importance of Data to Build Better Mobility Solutions

My last article concluded on the importance of data management to foster better mobility solutions. Let's follow up on the subject today.

This article focuses particularly on bus operators. The example is from MOBILITY ADO’s core operation, but all the concepts and processes described can easily be escalated and replicated to any transportation mode.

To better explain the importance of data, we will divide a mobility operator's activities into three phases: evaluation, design and operation. In the first phase, the operator or the public authority —often assisted by a specialized consultant— evaluates the demand through a deep understanding of current mobility patterns and future estimations based on urban and infrastructure projects. During the second phase —design— several service offers will be evaluated depending on how efficiently they will address the estimated demand. Finally, once implemented, the challenge is to execute the offer and ensure a high service level, respecting the programmed offer, also understood by the users as punctuality.

These three phases form a cycle. One can see some similarities with design thinking as this is a continuous process of testing  and learning, but there is one huge difference: temporality. It is very complicated —and expensive— to reach the same flexibility and agility on mobility services and infrastructures as exists in software development.

Evaluate

There is a great diversity of data sources available, each with different costs and levels of information. Before starting to assess the demand, it is important to define the objective in order to choose the right data source(s): Do we want to improve the service of an area that has just built a hospital or a school? Do we want to define the master plan for the extension of the subway over a 20-year period? Do we want to redesign the bus network from scratch?      Additional to the understanding of current and future needs, the offer in place needs to be understood in order to evaluate the current modal mix between transportation modes serving the market.

Among the data sources, we can mention search data from international planners, such as Google Transit or Moovit, that register every navigation search from A to B. Phone operators can track and sell data from moving devices that can be related with the socio-demographic category of their owners through the device type or model. Also, various IoT sensors, using mainly Wi-Fi or Bluetooth, can be used to track devices that are passing through a station or a terminal, as the device modems track records of all the access points (MAC address for instance) they have connected to, so one can deduce the “living” and “working” areas of these devices. Finally, the traditional in-field survey from the national statistical agency or mobility consultant is still very useful and precise for understanding users and their mobility patterns, as they can include a lot of qualitative information. For instance in Mexico City,  INEGI develops every 5 years the Household Origin-Destination Survey in the Metropolitan Zone of the Valley of Mexico, which is the basis for public policies on mobility and infrastructure.

Design

Once the needs are mapped out, the offer design can start. This is the phase of data modeling, data visualization, data optimization and artificial intelligence (AI):

●  Data modeling builds hypotheses and scenarios that can be virtually tested on the future needs. If we create a new route from A to B, mapping stations here or there, how efficient will it be to address needs? How will it impact existing modal shares? Will it properly address future needs?  

●  Data visualization attributes these scenarios to specific user interfaces, so that human beings can understand the impact of each scenario and move parameters and hypotheses around in order to improve possible solutions.

●  Data optimization is necessary to design the operation: it minimizes the investment in infrastructure, assets and people required to operate a defined offer.

Moreover, AI can exponentially improve models and algorithms, thus improving the design, from a macro perspective (the network design and connections) to the micro details (stops and station position on the street, fluidity of operations related to traffic lights, customer acquisition depending on accessibility and commodity, among others).

This process is a kind of virtual test and learn, a back and forth between mobility needs and mobility offer scenarios in terms of efficiency and costs. In general, it needs to be virtual because it would be too expensive to implement a transport line (like a beta version or an MVP) that could then be modified little by little gradually (like in development sprints) to reach the best solution, even if it is easier when no infrastructure is required. Put another way, a bus line can be tested if you have some available assets (buses and conductors) because it is less infrastructure-dependent, but it is impossible to build a metro line and then decide to change its design.

At the end of the design phase, the best scenarios are selected by the operators or public authorities. They include a theoretical offer: routes, stops location, stop durations, travel times, calendars, station and terminal design, number of assets/vehicles required, asset productivity, number of operators required, estimated traffic and affluence on each route segment.

Operate

Last, but not least, is the operation phase. This is the most important phase in terms of team dedication and time consumption because even if perfect agility would be ideal, infrastructure and asset barriers make it almost impossible today. In the end, the design of a subway network changes every century, every 30 years for bus lines, and so on.

At the infrastructure level, data is important to understand real- time events and prevent any issue. This is the world of IoT. Sensors measure roads and railway status (traffic and infrastructure conditions). Traffic lights are managed to pilot intersections in order to promote a BRT (bus rapid transit) over private cars. Sensors give real-time position, affluence and states of assets (vehicles). Infrastructure and asset data transmission rely on the network capacity to transfer mobile data in real time, so network connectivity is key. Also, the future development of autonomous vehicles, from cars to trains, will require huge investments in connectivity and the development of new communication protocols, such as V2V and V2I, vehicle-to-vehicle and vehicle-to-infrastructure communications, so that the assets and the infrastructure can exchange data and information to improve safety.

Regarding operations, real-time asset data allows the service to operate. Operators can see where the assets are, communicate with the field to make sure the theoretical plan is met, or record activities and affluence through embarked cameras. They can also add or quit assets so as to respect theoretical frequency and punctuality when an incident happens, plan predictive maintenance, and in the future, pilot autonomous assets from the operations center.

But real-time information is also really useful for customers who want to be aware of waiting times and trip times on every available channel: inside the vehicles, at stops and terminals, on websites and apps, through third-party platforms, among others. When autonomous vehicles come, it will be even possible to meet supply and demand in real time. In fact, measuring affluence can generate new services: it will be possible to add or quit vehicles in real time, thus modifying frequency and capacity.

Finally, the mobility-as-a-service platforms, public authorities and third-party stakeholders can consume theoretical and real-time data through APIs and standard open data frameworks, looking for integration between various services from different operators and transportation modes.

Building efficient mobility solutions is necessary to tackle the UN's global development goals. The mobility industry has existed for a long time, and its very traditional players need to adapt to the new challenges of digitalization, data and technology. The improvement of the service offer and the competitiveness of its actors is inseparable from the integration of technology and data.