Citizen-centric Design of Consumable Services for Smart Cities

For the general elaboration of the smart city setting, we first address that no universally used definition of the term smart city exists in the scientific literature before the importance of smart services within this setting is emphasized.

In general, a smart city is related to a set of approaches that use technologies to optimize existing or shape new city-related processes [9]. However, various terms (e.g., Wired, Intelligent, Digital, or Knowledge City) are used in the literature to describe different concepts of smart cities [16, 21], and definitions have also evolved over time, which might be explained through the effort of adapting them alongside the availability of novel ICT [62].

The main focus of the study at hand are the services resulting from the utilization of technologies within the smart city setting. Such smart services form a core building block of a smart city and are often provided by the city administration to be consumed by other entities of the city [4]. The goal is that the life of citizens, administrative processes, and the overall quality of life in the city improve through the incorporation of ICT [4, 71].

Within our citizen-centric development approach (cf. Section 3), we highlight that the citizens’ needs are of the highest priority when designing smart city services, as emphasized by other authors [62]. Moreover, a holistic view of the smart city is needed, in which citizens, technology, and the environment are treated as interacting concepts [49]. We, therefore, define a smart city in the following as a city, in which ICT is incorporated into the infrastructure and combined with existing technologies to generate benefits for the citizens in the form of smart services. Furthermore, the existing efforts in regard to smart city-related standardizations have to be acknowledged as a relevant factor for the design of smart city services. Such standards are created with the aim of providing guidelines for the consistent use of procedures across different smart city projects in order to ensure the interoperability of individual components [4, 35].

The notion of citizen-centricity is nothing new and was first introduced in the context of web-based government services [37]. During this time, citizen-centricity was about identifying the needs of prospective website users and adjusting the provided services accordingly [68]. In the following years, the term’s prevalence has developed alongside smart city research and often forms the foundation of related services. Nowadays, citizen-centric design is still about the assessment of the users’ needs in the first place, to then incorporate those needs into the final solution.

At this point, it has to be mentioned that closely related terms, like user-centric [8], human-centric [11], or people-centric [27] are often used in the literature instead of citizen-centric. We argue that these terms are used synonymously to describe the same concept since humans, people, users, and citizens can be conceptualized as having different roles of the same entity in the smart city context (as will be shown in Section 3 and Figure 1). Hence, the term citizen-centric is deemed more suitable for the given context and used throughout the remainder of this article while still implying the inclusion of the mentioned related terms. Additionally, researchers suggest that the participation of citizens and public servants during the design of smart city projects has a crucial influence on the successful implementation of the respective project [61]. Consequently, two closely related terms are often used in the scientific literature and need to be emphasized in this context, namely, citizen engagement and citizen participation.

Early on, it was recognized that new technological capabilities could increase public engagement through political platforms [52]. At that time, the use of the terms civic engagement and electronic democracy was more widespread for describing the new ways of communication between citizens and their respective city administrations. These first ideas of political platforms that enable electronic democracy were mostly techno-centric oriented, meaning that technological capabilities were the subject of interest. Nevertheless, the proposed benefit of such platforms persists since participation platforms are still proposed as a potential solution to foster civic engagement [61]. In more recent contributions, the term citizen engagement has been brought into focus, and the concept of participatory or collaborative governance [38], for which citizen participation is seen as a crucial input [22], has mostly replaced electronic democracy.

Nowadays, the focus of citizen centricity and related concepts have shifted from a technological capability to a social necessity [11, 20]. This is partially the case because the elicitation of citizen engagement poses a major challenge to today’s smart cities [13]. Furthermore, its importance for smart city development has been recognized in the literature [17], which can be explained through increased democratic participation by the population, leading to higher approval rates, trust, and also quality of life among citizens [22]. Other contributions even suggest that citizen engagement is linked to the development of so-called sustainable-oriented innovations (SOIs) [64].

Finally, the concept of open data, and specifically, open government data, is introduced in the context of citizen centricity and smart city development due to the numerous effects it is said to have on the related terms presented in this section. Before these effects are discussed, it is important to mention the different terminologies used in the literature that describe different elements of the same concept. Among them are open data [2] in combination with open data initiatives [51, 53, 69], open data portals [55], and open government data [1, 39, 72]. Open data includes all structured data that is publicly available in the respective city, while open government data describes all data made publicly available by the government or municipality. Open data platforms are essential for the management and publication of open data within cities, and all efforts towards establishing or improving such platforms form open data initiatives. Since open data platforms are usually operated by the government and can include data from open as well as governmental sources, it is hard to distinguish between open data and open government data in this context.

More importantly, open data initiatives are generally considered to be an essential building block of developing smart cities [51, 72]. The reasons for these effects are diverse, stating that open data can foster innovations [53, 69], citizen engagement [7], citizen participation [62], and even promote trust building between city municipalities and the respective citizens [51]. Increased trust in this relationship is especially important considering the movement in the opposite direction, resulting in citizens trusting the governmental institutions less and less [46]. Hence, another focus of most recent contributions has been the challenges connected to the publication of open data [55] and the assessment of the transparency of current open data initiatives [39].

A citizen developer is considered a developer with little to no experience in software engineering or coding. Yet, they use so-called low-code or no-code platforms designed to enable fast, easy, and cost-efficient application development [57]. The idea behind citizen development originates from the problem of shadow IT systems within organizations where employees use applications not provided by the organization to compensate for specific shortcomings [24]. Consequently, employees have been encouraged to develop such workarounds through in-house low- or no-code platforms in order to distribute resulting applications within the whole organization. A similar approach is present in the smart city literature, namely, in the form of participation platforms [61]. In this study, we distinguish between citizens that use smart city services and those that produce them, although both roles can be occupied by the same entity. In the following, the first is referred to as citizen as consumer while the latter is referred to as citizen as developer.

The same notion of a citizen acting as both producer and consumer can be found in [1]. The authors use the term prosumer to describe the possibility of citizens both consuming data services and, at the same time, serving as providers of data for these services through their smartphones with the goal of generating citizen-centric applications. Nevertheless, it has to be mentioned that the term prosumer was originally coined by Alvin Toffler in 1980 [63] and used to describe the shift of companies outsourcing tasks to their customers (e.g., self-checkout systems and fast food restaurants). Later, the same approach was applied during the advancement of the web, where consumers suddenly became responsible for creating and maintaining the content they consume (e.g., YouTube and Wikipedia). Consequently, the term user-generated content (UGC) quickly established itself in this context as a competing concept to the prosumer.

For our study, we deemed the term citizen developer to be most suitable in the smart city context for representing the underlying roles and relationships between them. Keeping the origin of the term in mind, we assume that a respective smart city is equivalent to the organizational environment mentioned above. Under these conditions, it can be argued that, instead of employees, citizens are looking for applications other than the ones provided by the city administration to compensate for specific shortcomings in the city. Such shortcomings could for example be the lack of a suitable platform to develop city tours. In this case, the city administration should aim at providing citizens with a development platform for the realization of their ideas in a low-code or no-code fashion (cf. [48]). The resulting applications and services can then be offered within the smart city environment to benefit all citizens. Hence, our interpretation of the term citizen developer not only ensures the incorporation of citizens’ needs during the design of smart city services but also accounts for the responsibility of city administrations to ensure citizen-centricity.

The aim of the citizen development approach is to ease the creation of smart city services for people with little to no software engineering skills. This aim requires a comprehensive specification of people’s knowledge so it can be understood by humans and processed by machines. For this purpose, the use of computer-aided conceptual modeling is proposed, which implies that a modeling tool exists to create, digitally save, and process the models.

Conceptual models are abstractions of a domain under study, focusing on the representation of concepts and their relations within models [40]. Concepts form generalizations of objects, which are represented through their properties. Reducing concepts to these essential properties describes the abstraction process directed towards a specific purpose. Thereby, abstraction reduces the complexity of the system under study to make it manageable [40].

In computer science, conceptual models are used to describe information systems [40] with the goal of fostering communication and understanding between multiple stakeholders. To serve this goal, models are created as instances of modeling languages, which define the notation, syntax, and semantics of the models [30]. The benefit of using such modeling languages is that everyone familiar with their definition understands the created models. Furthermore, it is argued that the value of models is increased if they can be automatically processed by tools to put the captured knowledge to work [30]. Such capabilities distinguish the modeling language from the modeling method, as the latter also includes automated processing and guidelines for creating models in addition to the language itself.

Modeling methods should be designed in such a way that they can be interpreted by both humans and machines in the form of computers or other technologies. To make models readable for humans, they must be visualized through their notation. One way of doing so is diagrammatic visualizations, which have the benefit of encoding spatial information in the resulting models [25]. Additionally, computer-aided diagrammatic modeling can further facilitate human understanding if semantic-rich symbols are used to represent the concepts of the domain [44].

The interpretation of models by machines is enabled through the respective metamodel, which describes the underlying architecture of modeling methods in a machine-processable way so that every model instance can be interpreted [28, 30]. Metamodels are created for a specific purpose by applying abstraction to capture the main concepts relevant when creating models in a given domain. These abstractions are then enriched with semantic-rich notations to also foster the human understanding of the metamodel.

Certain technologies have to be offered and maintained by the city municipality to allow the creation of services in line with the proposed citizen development approach. In this context, technologies are not necessarily physical, meaning they can encompass any kind of applied knowledge used for a specific purpose [18]. Based on this understanding, two technologies relevant to our approach are introduced: microservices and multi-sided platforms. Additionally, we introduce the Scene2Model tool, which forms the foundation for the multi-stakeholder service model design proposed in Section 3.3.

Mircroservices. The development of ICT-based solutions in the smart city is closely associated with IoT, and open data [2, 71]. The resulting applications pose complex challenges for software systems as the interoperability of different technologies has to be ensured [35]. One way to tackle this complexity is the application of web services in the form of a microservice architecture [36]. The functionalities of a software system based on such an architecture are separated into decoupled services, each focusing on an independent functionality [19]. Every microservice is written in its own programming language and possesses its own technology stack. As a result, the microservice architecture reduces the dependency within the software system and supports individual deployment, maintenance, and replacement. Each service runs independently, and communication between services utilizes light-weighted communication protocols (e.g., REST-like interfaces). The workflows of the software system are then executed through the connection of different microservices. Two ways of accomplishing this execution are orchestration, and choreography [19]. Orchestration means that a centralized service exists, which handles the execution of the workflow. On the other hand, such a centralized service does not exist in choreography, and the workflow is embedded into the microservices. We assume that different services in a smart city are provided by multiple stakeholders, like other citizens, businesses, or the municipality (cf. Section 3). Stakeholders should thus be able to define their own services by orchestrating existing ones, as it is often done in microservice architectures. The logic of the created services is then described through conceptual models, which are executed in a microservice environment (cf. [66, 67]).

Multi-sided Platforms. Following the notion of citizen development, an ecosystem must be in place to enable citizen-centric service creation. The concept of multi-sided platforms is introduced in this context, as it offers a suitable orientation for a foundational architecture that supports such an ecosystem. Traditionally, one-sided platforms act as intermediaries that focus on connecting users with providers of specific products or services without an actual interaction between the two parties (e.g., online retailers or streaming services). In contrast to one-sided platforms, which only serve one specific group of users, two-sided platforms are based on generating value through interactions of users with compatible needs belonging to different user groups (e.g., ride-sharing or dating platforms). Multiple dependencies exist between the involved parties in an ecosystem like the latter, as everyone needs to participate in creating value for the other parties [56]. Namely, the provision of offerings, the satisfaction of user needs, and matching the two in a suitable way through an intermediary platform form the underlying prerequisites that have to be fulfilled for a functional multi-sided ecosystem [58]. The platform provider has a special role in this context. They must mediate the interaction between the involved parties and ensure a balance between demand and offerings to foster the value exchange [12]. A foundational architecture for the proposed citizen development approach is oriented on multi-sided platforms, therefore, has to contain at least a platform provider that acts as an intermediary between the involved parties. Notably, these parties of consumers trying to satisfy their needs and providers offering services on the platform can be the same entity, as in the case of AirBnB.

Design Thinking and the Scene2Model Tool. The design and creation of services in complex environments, such as smart cities, often require the collaboration of multiple stakeholders. To address the challenges present during such collaborations, design thinking has been shown as an effective approach (cf. [14]). One way of utilizing this approach are physical workshops that bring together domain experts and employ creative methods to address communication and coordination issues resulting from the presence of diverse stakeholder interests [59]. The results of such workshops aim at capturing the externalized knowledge of the workshop’s participants in the form of physical representation, e.g., through paper figures on a table. Scene2Model addresses the challenge of capturing and distributing the externalized knowledge by offering a physical and digital modeling environment that allows an automated digitalization of physical storyboards created during the workshops [41, 42]. The physical storyboard approach uses paper figures to represent stories of users designed in the workshop and is based on SAP Scenes^TM.^* These storyboards are then automatically transformed into digital models, which contain not only the graphical representation of the physical scenes but also each of the paper figures as its own object that can be enriched with information or changed individually [47]. The automatic transformation is an essential feature, as it removes the cumbersome task of creating documentation (e.g., the models) for users, allowing them to focus on creating and representing the idea in the workshop. Moreover, the knowledge represented in this way is computer-processable and, therefore, improves its utilization by easing the adding of information to the storyboards as a whole or to its individual concepts.

The architecture for our proposed citizen development system must account for the different entities and technology-enabled interactions that exist within (cf. Section 2.6.2). For this purpose, the interdependence model from our previous work was extended to account for all entities and components relevant to the functionality of an interconnected system in which value is created through the exchange of resources in the form of knowledge or services [3, 65]. In this context, interactions enabled by technologies play a crucial role [45]. Hence, different entities, digital as well as physical technologies, and their various interactions are involved and therefore have to be considered in this value-creating system. The resulting abstraction in the form of a conceptualization is displayed in Figure 1.

The smart city infrastructure contains consumable services, services models, its environment, and technologies, with utility services being part of the latter. In reference to Section 2.5, the city infrastructure and the respective city administration governing it can be defined as the platform provider, that enables the interaction of users with compatible needs.

The environment component represents the physical attributes and properties of the respective city, which include static attributes (e.g., location of roads and buildings) and dynamic properties (e.g., weather, live data on traffic or public transport). The term technology has been defined in various ways, depending on the historical period (e.g., prehistoric and modern), the underlying perspective, and several other factors. To be suited for the context of this study, the high-level definition presented in [18] is adapted, which entails that technology stems from the application of derived knowledge that is used for a specific purpose while the output does not necessarily have to be something physical. The technology component thus encompasses not only physical objects but also digital applications offered within the smart city infrastructure. One example would be an endpoint that provides open data access (cf. Section 2.2), which can support the creation of services through citizen developers [53, 69].

Nevertheless, technology cannot directly interact with the physical environment, which is why a technology-enabled interaction must be established. This interaction is represented in Figure 1 through the relations between environment and technology, namely, actuation and sensing. These ways of interaction can be provided by incorporating CPS into smart cities [33], which allows the physical components to be enhanced and made accessible to the services.

In this architecture, the entity citizen can interact with the smart city infrastructure by taking on different roles. As a result, the term citizen describes not only a person living in the respective city but also all entities that interact with its infrastructure through varying roles, depending on their respective goal. Citizens either consume already available services, which lets them take the role citizen as a consumer, or they provide their domain knowledge for the creation of services, which lets them take the role of citizen as a developer.

Consumers interact with the smart city infrastructure through the consumption of consumable services to achieve a certain goal. Although consumers don’t provide services themselves, they still generate value for the system by paying a price or creating demand, which attracts new providers. Furthermore, the consumers trying to satisfy their needs form another prerequisite defined in the context of multi-sided platforms (cf. Section 2.5).

The interaction as citizen developer happens through the creation of consumable services, and their domain knowledge qualifies them to solve problems they experience themselves. In this context, the citizen development approach employed in this study (cf. Section 2.3) supports citizens in creating solutions in a low-code fashion without the need for sophisticated software development skills. This is achieved through the application of domain-specific modeling methods that provide the means of defining the solution in an executable way. Citizen developers then share their consumable services with other participants on the platform and thereby form the second party of providers offering services (cf. Section 2.5). Their motivation can range from altruistic acts of helping other citizens to monetary compensations they receive from consumers of their services or through agreements with the municipality.

To reduce the complexity of the creation process and support citizen development even further, a consumable service makes use of different available utility services and other technologies from the smart city with the goal of combining existing functionalities. An example of a utility service is the offering of a payment service, which citizen developers can easily reuse and integrate into their respective design of a service. Consequently, citizens or the city municipality can act as a utility service provider, which forms an extension to the roles introduced in our previous work (cf. [48]). The motivation for providing utility services can again be monetary value for citizens, as they can offer their services to a broader audience on a pay-per-use basis. Municipalities or governments, on the other hand, can stimulate the system by providing their own foundation of utility services as part of an open government initiative while also offering incentives for every contribution to the resulting multi-sided ecosystem.

The reusability of existing functionalities in the form of utility services is enabled through their microservice architecture as described in Section 2.5. That implies that the utility services are developed independently from each other and run on their own, requiring utility service providers to have software engineering skills, in contrast to the citizen developers. To enable such a combination of services, we follow an orchestration approach [19], where a central unit coordinates all microservices. Within our conceptualization, the orchestrator is represented by a consumable service, which is defined through an individual diagrammatic model (cf. Section 2.4) that can be directly executed within the proposed architecture [67]. Each citizen-developed service forms its own orchestrator of selected utility services which require an interface that is compatible with the citizen-developer platform to enable seamless integration.

Service models form the final component within Figure 1, which represent the abstract specification of citizen developers’ domain knowledge lying at the heart of each consumable service. As visualized in Figure 1, the service model alone cannot be executed or consumed as it only forms the specification of the domain knowledge. Hence, it must be configured within the modeling environment to allocate the right interface of utility services and other technologies. This enhanced specification of a service must then be adapted to the respective city infrastructure in order to transform a service model into an executable instance of a consumable service. One essential part of such a transformation is the incorporation of suitable utility services and other technologies to provide the functionalities needed for the execution.

To allow for a separation of the service model and its executable instance, the defined tasks are not directly linked to the interfaces, as this makes the services dependent on the respective smart city technology. The additional abstraction between the technology interfaces and the service model through capabilities, as displayed on a conceptual level in Figure 2, is thus proposed. Capabilities represent an abstraction of the tasks that are needed to execute services within a given infrastructure. In the case of a specific executable instance, the capabilities are allocated to the required technology interfaces of the city infrastructure. Furthermore, if the technology within a smart city changes, a limited number of interfaces have to be reconfigured so that consumable services can be executed within the new infrastructure. This adaptability requires that the technologies are based on a microservice architecture that separates the abilities of the smart city into independent and reusable components offered through a defined interface.

When describing a specific scenario within the citizen development architecture, the relevant components of the smart city infrastructure have to be instantiated. An instantiation of the drone tour guide case is displayed in Figure 3, which is reduced to exemplary environment and technology components.

The drone tour guide scenario takes place in the smart city environment of Vienna. The environment contains both static (e.g., Vienna Attractions and Vienna Map) as well as dynamic properties (e.g., Weather) and interacts with the present technologies through actuation and sensing. This technology-based interaction is enabled through both static and dynamic CPSs organized in a microservice architecture that are incorporated into the city infrastructure. Static CPSs like sensor networks are strategically placed in defined locations. In contrast, dynamic CPSs, such as flying drones, gather and use information on weather and traffic conditions while moving through the city.

Additionally, the technology component also contains digital applications that are represented in Figure 3 by selected utility services considered helpful in the context of the scenario (e.g., Payment Service, Routing Service and Live Weather Data). Such services are provided by citizens acting as utility service providers. As elaborated before, this role can be occupied by both citizens and organizations which are motivated through monetary incentives, or the city municipality offers the services as part of a government-supported initiative (cf. Section 3.1).

Assuming that the described environment and technology in the infrastructure of Vienna are in place, citizens can interact with the smart city as citizen developers by creating service models that are later transformed into consumable services to satisfy a consumer’s needs. Within the presented instance of the drone scenario, this can be citizens of Vienna that decide to use their knowledge about the city to create a tour designed for a specific purpose. The citizen developer benefits from the microservice architecture of the utility services, which enables easy integration of different capabilities provided by the available technologies into the resulting consumable service (cf. Figure 2).

An exemplary instantiation of the composed consumable service in the context of the tour guide scenario is represented in Figure 4. In this representation, the citizen utilizes the benefits of the microservice architecture by designing the respective tour as a process within a dedicated modeling environment. The different tasks contained in the tour require specific capabilities, like the selection of a tour from an existing database. To execute the defined tour in a real environment, these capabilities must be allocated to the technologies present in the respective smart city environment. This allocation process is enabled through interfaces, which abstract the functionality that the technology components (e.g., users’ smartphone, drone as a tour guide, and available smart services) offer to the environment. For example, the tour selection needs to retrieve a list of all city tours and provide them to the interface of the user’s smartphone.

Through the allocation of capabilities to concrete interfaces, the drone tour model from the example is transformed into a consumable service within the Vienna infrastructure that can be used by citizens acting as consumers. Although this service model is specified for a drone, the microservice architecture allows for the reallocation of utility services so that the tour can be adapted for the execution by different technologies available in the infrastructure.

Previous elaborations on the smart city setting (see Section 2.1) did not cover the fact that multiple stakeholders are usually involved during the design implementation of smart city services, which further complicates the process [5]. Relevant stakeholders could potentially include city planners and technicians to implement necessary infrastructure changes, citizens influenced by the service, and representatives of the city municipality. Furthermore, citizen developers and utility service providers should also be considered relevant stakeholders in the smart city context.

We argue that the initial phase of the innovation process and identification of requirements for future service models can be supported through both creative workshops and conceptual modeling. The goal is to identify citizens’ problems and needs and capture them so that they can be understood and shared with involved stakeholders. In Figure 5, an overview of the most relevant concepts in our multi-stakeholder design approach is visualized.

The central component of our idea is the use of citizen workshops to bring multiple stakeholders together and allow the exchange of ideas in a comprehensive manner. Citizen workshops are inspired by creative workshops used in design thinking, whereas the name itself highlights the crucial role of citizens participating in the workshops. The physical design thinking workshops are used within our approach to facilitate the co-creation of multiple stakeholders, and the resulting knowledge in the form of tangible models can be digitized as digital conceptual models through the Scene2Model tool (cf. Section 2.5). Therefore, the knowledge itself is created through creative methods in the workshop, and Scene2Model is used to capture this knowledge in digital models to support its utilization and distribution.

Digital conceptual models are a known tool used for describing information in a comprehensible way and are thus suitable to serve as the medium for knowledge exchange between stakeholders (cf. Section 2.4). In this article, the focus lies on diagrammatic models since graphical visualization significantly improves comprehensibility for humans.

The digital models resulting from a workshop can thus be used to communicate innovative ideas between stakeholders by sharing them directly with other citizens or evaluating them in a collective intelligence environment. This entails sharing enhanced models with a large group of people to gather feedback, for example, through social media. The modeling environment of Scene2Model enables the automated publication on such a platform and later gathers the feedback, automatically processes it, and saves it with the model. In this context, sentiment analysis can be used to analyze the mood of people replying to social media posts and assess if the general idea is well received by a broader audience [70]. The created innovative idea and the evaluated digital conceptual model (representing the results from the citizen workshop) can then be used as the basis for the development of service models as described in Section 3.1.