Engaging non-scientists in scientific data collection and research is known as citizen science . It has also been defined as scientific citizenship, foregrounding the necessity of opening up science and science-policy processes to the public. It has been widely discussed as the public participation in science and communication projects . Today, a range of concepts have emerged referring to citizen science and participation, such as volunteered geographic information, crowdsourcing geospatial data, people-centric sensing, participatory sensing, and mobile crowdsourcing. One of them, the citizen observatory (CO), has received a great deal of interest in terms of its applications and research challenges.
What are COs? Where and how are they being used? Is it technologically feasible and acceptable that citizens are engaged in societal challenges via COs? Here I discuss these questions while showing how HCI technology helps not only in engaging citizens but also in creating complex human-data interaction processes. Citizens, then, are more than data collectors—they are innovators and co-designers of data-intensive services. Based on the survey we conducted , we highlight some challenges related to citizens’ engagement in COs and the underlying human-data interaction model.
In a broad sense, a CO is defined as the citizens’ observations, perceptions, feelings, expectations, and comments on concerns that they report. A CO is a platform for interaction between citizens and scientists for the engagement of a very wide community of users, for example, in a societal challenge. The range of applications includes health tracking, city and transportation management, crisis management and response, peacekeeping, sustainable development, as well as homeland security. It can be used to address any societal challenge. In 2016 we analyzed more than 100 citizen observatories with the goal of identifying HCI design challenges, including citizen-engagement techniques and the interactive software technology used for CO design and development.
- The monitoring of global concerns (e.g., Galaxy Zoo, Springwatch, Globe at Night)
- Disaster tracking, crisis response, and prediction (e.g., iShake, Did You Feel It?)
- City management and decision support concerning traffic, parking, bicycle routes, environmental conditions, and citizen perceptions of municipal services (e.g., FixMyStreet, MITOS, CyclePhilly)
- Insect, bird, butterfly, and sea-species monitoring (e.g., eBird, Great SunFlower, iBats)
- Biodiversity management observatories on flora, forests, and the biosphere (e.g., Leaf Watch, COWEB, Plant Watch)
- Water, stream, snow, and sea observatories, including precipitation and water properties (e.g., Brooklyn Atlantis, CoCoRaHS, WeSenseIt)
- Air and spectrum monitoring such as air, noise, sound, and radiation (e.g., ekoNET, CitiSense)
- Various other CO-development platforms that have been proposed, such as SETI Live, Ushahidi, and Fold.it.
We have also identified some of the practices used to engage citizens in these COs (Table 1). An analysis of the user experiences with the current citizen observatories can help to better elicit the attributes that motivate and engage citizens. These can be cognitive, economic, or social.
|Table 1. Examples of modern citizen observatories .|
Human-data interaction (HDI) is an emerging field of research grounded in the tradition of HCI and some related disciplines, such as media and information studies. The term, coined by researchers from the MIT Media Lab, generally refers to the collection, storage, analysis, and use of personal data. This includes data from multiple sources, private and personal data, open data, and data from networked sensors. All this data can be put together or linked via humans. A CO forms a data-intensive, sociotechnical, highly interactive system. It supports intensive human-data interaction. Citizens, scientists, and stakeholders interact with the data and with each other while discussing problems and contributing to the co-creation of solutions. As a human-data interaction platform, the user interfaces of a CO should then support four human activities:
- Engage and motivate citizens using crowdsourcing and social media technologies. The gamification of such technologies can help to solicit contributions from a wider community of citizens, especially the existing online ones. How can gamification techniques and gamified design be used to engage citizens?
- Connect using the Internet, the Internet of Things, and networked sensors that can allow physical objects to be remotely controlled by citizens via COs. The concept of tangible user interfaces can be applied. For example, physical objects such as trees, forests, and lakes can be enhanced with sensors and considered tangible user interfaces. Humans can interact directly with the data being collected using these new forms of user interfaces. Will humans be able to use these enhanced physical objects with sensors as a new way of interacting with the digital world? This is a key question for HCI researchers.
- Collect, store, and deliver data easily and quickly. Mobile phones and wearable devices can be enhanced by embedded sensors to automatically gather, exchange, and disseminate a wide range of data, including images and videos.
- Extract and collaboratively discover patterns. Existing information visualizations need to be re-investigated to understand the context in which the data collected by citizens is used, manipulated, and traded for pattern co-creation.
These activities form the foundation of a model for human-data interaction in COs, the 2CE Model: Connect and Engage, Collect and Extract. The following are the major technological motivations for this model.
First, most COs today use a service-oriented architecture (SOA) as a development and deployment model. COs are then seen as a large composition of data-intensive services. A user interface can also be designed as a service. A user interface as a service (UlaaS) is introduced here to refer to a specific device, a platform, and a context of use . Various UIsaaS can be related to the same CO’s service. This would mean there is no need to download and install any heavy UI toolkits. Citizens are provided with a UlaaS that they may select and adopt. It is accessible most often via the Internet, to collect and interact with data. However, innovative UIsaaS need to be created to support the core activities that form the 2CE model. Should we provide all citizens with the same UIaaS to interact with data, or should we develop a UIsaaS that citizens can customize and use? How do we design such mosaic UIsaaS? Are existing HCI evaluation and user research techniques appropriate for UIsaaS? These are questions needing answers.
Second, portable microsensors are making observations accessible to the common citizen. However, citizen experiences in observing and understanding environmental concerns, particularly in reporting and commenting on them, need to be enhanced. Participatory sensing is making human-data interaction more distributed. Individuals acting in groups may use their personal smart mobile devices to systematically explore data about their lives and social environment. How can enhanced mobile phones with sensors engage millions of people? How can this raise awareness of environmental issues, for example, by supporting the lifelong education of citizens?
Third, we strongly argue that it is not enough to be able to collect, store, and visualize large amounts of data. Citizen scientists require innovative visual-mining techniques to connect data from different sources while extracting insights and patterns. Crowdsensing technologies are being used to capture data using enhanced smartphones (Google Nexus, iPhone), sensor-embedded gaming systems (Xbox Kinect, Wii), and in-vehicle sensor devices (OBD-II and GPS). Furthermore, a CO should provide the semantic meanings of data when citizens interact with it. For example, users of earth observations have a wide range of data requirements that depend on their specific applications. Some users need both raw datasets (of directly observed phenomena and derived forecasts and products), while others utilize only a particular type of dataset. New non-intrusive user interfaces need to be created to support these citizens. They should mediate the collaboration between these citizens and the scientists.
The following are the three main challenges we identified from the more than 100 COs we investigated in our study.
The first challenge is the citizen repository and the user interface architecture and engineering to facilitate their adaptation, usability, and accessibility. COs are highly interactive systems. The UIs in current COs range from traditional GUIs, mobiles, and tablets to specialized devices, such as sensors connected to a UI for monitoring and enhanced everyday physical objects. The difficulties of designing such a mosaic of UIs are exacerbated by the huge amount of possible combinations of UI development kits (SDKs) underlying this diversity of users, looks and feels (LFs), UI guidelines and patterns (GPs), and operating systems (OSes). Designing a UI for SDKxLFxGPxOS will require an armada of HCI designers and service developers. How do we facilitate the design and development of such mosaics of user interfaces?
One architectural pattern is to design the UI and CO as a service and to incorporate experiences of COs into the design loop  (Figure 1). The CO is a combination of reusable services. A service is a building block that performs a specific function—a function might be to retrieve data collected by a citizen. Figure 1 portrays a CO as a collection of geographically distributed Web, mobile, and cyber-physical services. Context-aware and platform-dependent services are combined and accessible via distributed user interfaces. CO as an interactive service is built on the top SOA standardized protocols, platform independence, well-defined interfaces, and tool support, which allow legacy systems and data sources to be easily integrated and secured, and the user interfaces to be easily adaptable to the context of use. Various technologies are being developed to facilitate the seamless integration of physical objects and software required by COs. Emerging Sensor Web Enablement standards and GeoPackage formats are making any physical object accessible via the Internet of Things.
|Figure 1. An architectural pattern for designing COs and UIs as a service.|
The second challenge is related to the essence of COs: the level and strategies of citizen engagement. This requires us to identify citizen-engagement patterns and to design UIs that implement those patterns. Citizen engagement has evolved from being aware in the 1960s, incorporation of local perspectives in the 1970s, recognition of local knowledge in the 1980s, participation as a norm as part of the sustainable development agenda of the 1990s, and e-participation governance in the 2000s . Based on our study, we proposed a model that distinguishes six levels of citizen engagement:
- Brokers for data that they did not collect (citizens may use social media to share data they found with their friends and scientists)
- Data collectors, who use digital devices to collect and send data that they are requested to monitor
- Data investigators, who help professionals and scientists to find and assess the relevance of the data that they need
- Data analysts, who can comprehend large datasets and can extract insights and patterns
- Co-designers of services—data analysts who also participate in data-intensive service design, for example, in the design of a service to predict pollution in a lake using data about the water quality and species
- Innovators, who contribute to the co-creation of the societal changes and added business value that result from the development and usage of the COs.
To communicate these levels of citizen participation, citizen-data interaction, and citizen experiences, we could use personas (Table 2). The challenge then is to identify patterns of interactions and to design UIs that support each of these levels of engagement.
|Table 2. A persona-based model for defining citizen engagement.|
The third challenge is related to the fact that the user interfaces for COs should not only help every citizen to collect data, but also ensure that the data being collected is and will be accessible, usable, and useful. How should a user interface help citizens to collect the right data in the right way? It is well known in many areas, including COs, that there are two sets of attributes that influence the quality of the data being collected, and in particular, people’s preferences and decisions: 1) the degree of usability, trustfulness, and accessibility, which is influenced by the past experiences of the data provider/receiver and by the other community members who participated in data collection and dissemination; and 2) the stakeholders’ assessment of the risk/benefit trade-offs associated with data usage. This requires measures of privacy, accountability, and usefulness.
Figure 2 gives an overview of a quality model for human-data interaction in COs that we are developing. It depicts and compares the six quality attributes of COs and their foci, such as the questions what, how, where, who, when, and why, which end users and other stakeholders ask about the data and the way in which they interact with it. The model is bounded in a 6 x 6 matrix with six quality attributes as columns and the six W-questions as rows. The classifications are expressed by the cells, that is, the intersection between the six Qs and the six Ws. Each cell should provide an answer to a question from the perspective of a specific quality attribute and the experience of a citizen.
|Figure 2. The Six Ws/Qs Quality Model for Human-Data Interaction.|
The following scenario illustrates how correlations among different datasets can affect the quality of the human-data interactions and influence citizen experiences.
Suppose that an external source releases a study about the relationship between pollutants and the diseases of lake fish. By analyzing environmental data collected by a municipality and by citizens, and comparing that data with study results, an insurer could decide to increase the risk associated with citizens living in a polluted area and re-compute their policies. The association of environmental data from a CO with other data sources can be exploited for inferring human-rights/privacy-sensitive concerns. Suppose that someone has access to data recording the medical histories of the community. He or she might then link such data with pollution data from the CO and violate citizens’ privacy (adapted from ).
This model is derived from analogous structures grounded in the disciplines of architecture, construction, engineering, and manufacturing. The model classifies the artifacts created during the process of designing and engineering complex products (e.g., buildings or airplanes). The model considers data as a design artifact being created, stored, used, and destroyed, enabling focused concentration on selected aspects of the quality of data from a human perspective, without losing a sense of the context of use of the entire CO.
Finally, there is no doubt that COs are a powerful platform for collecting data and for engaging citizens in solving societal problems . Their long-term success requires addressing the following HCI issues:
- The architecture of COs to facilitate the seamless integration of the various forms of UIs and data, as well as their adaptation to the huge diversity of citizens. How can these UIs be designed as a service accessible from everywhere, anytime, and for everyone?
- The field studies that identify success and failure stories on how COs are being designed and used. Which HCI design methods are actually used or can be used?
- The use of social-media crowdsourcing combined with gamification strategies that empower millions of users to be engaged not just as data collectors, but also in the design and innovation process to turn data into new, innovative services and solutions.
Ahmed Seffah is a professor of HCI and human-centered software engineering. His interests include design patterns, theories, and practices, especially for cyber-physical systems and their applications for societal challenges such as sustainable development, homeland security, and crisis management. He has authored six books on the bridges between HCI design, design science research, and software engineering, and on UX/quality-in-use measurement. firstname.lastname@example.org
©2018 ACM 1072-5520/18/03 $15.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2018 ACM, Inc.