How do students use artificial intelligence tools in coursework projects when given the liberty to do so, with the only requirement of documenting how, where and why? We describe experiences with two groups of undergraduates in courses related to serious game authoring and human-computer interaction, both carried out in the second semester of 2023. In the serious games course, students were given the option of following a teacher-developed methodology for generating graphical assets for their serious games using a set of generative AI tools. This methodology was explained in the class but not hands on lab was carried out. In the interaction course, students were free to choose which AI tools to use when designing their system or in the development of their project documentation. Despite the limited number of participants (41 in total) we can see very different views and degrees of involvement: while some tried to use AI for as many tasks as possible, others considered that the learning curve for those tools was too steep to be worthwhile. Both experiences included a free-text survey at the end, and taken together, provide insights into how both supervised and unsupervised generative AI use could impact undergraduate projects in similar subjects. In addition to describing how students chose to use the tools, and the main takeaways from their survey response, we also discuss some of the ethical aspects about the access to the tools and what should be the minimal conditions to be met to allow the equitable use of AI in the classroom.

Serious games have several advantages for learning over more traditional activities. One of those advantages is that being highly interactive, they provide a firsthand simulated experience to the users. Serious games can also generate interaction data that can be used to assess learning, both for evaluation and to provide timely feedback for learners or even adapt aspects of the game on the fly. However, this potential is currently far from realized because doing so requires significant investment and expertise as compared to more traditional educational activities. Additionally, game learning analytics continue to be a costly and fragile process. We propose a combination of tools that greatly reduces the associated costs of building serious games with meaningful analytics simplifying the validation of those games and their deployment in real settings. We consider that this can be a key step to building predictive assessment avoiding the need for disruptive external assessment thus bootstrapping serious games to recognize learning through analytics.

The serious game industry has been a subject of interest for the educational community in the last few years. However, despite their potential benefits when compared to traditional e-learning, use of serious games in education has been hampered by difficulties, including that of deploying games in multi-platform settings and leveraging data generated by their players to improve the learning outcomes of those players. This increasing uptake requires to simplify serious games application for both teachers and developers. This paper proposes a standards-based approach to improve serious games deployment and evaluation. Building on existing and upcoming standards will allow games to be seamlessly integrated into the learning management systems that are already present in most higher education institutions. We present initial experiments that put the proposed approach into practice using two specific tools: uAdventure and SIMVA.

Data science applications in education are quickly proliferating, partially due to the use of LMSs and MOOCs. However, the application of data science techniques in the validation and deployment of serious games is still scarce. Among other reasons, obtaining and communicating useful information from the varied interaction data captured from serious games requires specific data analysis and visualization techniques that are out of reach of most non-experts. To mitigate this lack of application of data science techniques in the field of serious games, we present T-Mon, a monitor of traces for the xAPI-SG standard. T-Mon offers a default set of analysis and visualizations for serious game interaction data that follows this standard, with no other configuration required. The information reported by T-Mon provides an overview of the game interaction data collected, bringing analysis and visualizations closer to non-experts and simplifying the application of serious games.

This is an Innovative Practice Full Paper. Serious Game learning analytics can provide insight to improve both the learning process and the lifecycle of games. Due to the complexity and diversity of topics involved, from game design to implementation to tracing the user, teaching game learning analytics to engineering students is challenging. We have created a teaching approach built on an integrated set of tools to minimize the boilerplate setup and configuration typically required when building game learning analytics from disparate modules. Our approach relies on the combination of an authoring environment that eases the creation of adventure and location-based games, a standards-based interaction tracker, and a cloud-based analytics framework. In this environment, students can design and implement serious games with associated analytics models from the very beginning, allowing them to experiment from early on with analytics to improve their games' lifecycles. We have piloted this approach in two university courses for two years, where students prototyped serious games and then used analytics to understand how their users interacted with their games. The use of analytics helped our students to reflect on and refine their designs. While our approach can be applied with any tools, our authoring environment and analytics tools are available as open-source projects to facilitate development, prototyping, or testing of games with analytics.

The perceived complexity of many engineering courses is partly due to difficulty visualizing or interacting with the underlying mathematics. Traditional game-authoring environments lack STEM-related support, making them ill-suited to build serious games to explore these concepts. To address this issue, we have integrated an external math tool into the Unity game-authoring environment, and tested it by creating a game to teach basic concepts of mobile communications to engineering students. Players of the game had to fine-tune parameters in a simulated spectrum analyzer. We describe how we integrated STEM support for the generic authoring platform that we used, and present results from a pilot test of our game. We believe that this approach is extensible to a wide range of STEM disciplines.

Despite the proven benefits of Serious Games when compared to traditional e-learning, uptake of game-based learning in mainstream education is still very low. Increasing uptake requires making suitable games easier to deploy as an activity type in existing e-learning platforms, without compromising game quality or increase development cost; and leveraging data generated by players to improve their learning outcomes. We briefly discuss available e-learning standards to address these issues and describe our current standards-based approach to both game deployment and evaluation (including learning analytics).

The e-UCM group is working to simplify the processes of both creation and scientific validation of educational adventure games, by making the entire process of game creation, data acquisition, and analysis more transparent and integrated. This paper presents the integration of the uAdventure authoring tool with SIMVA, a tool to simplify validation experiments, thus encompassing the full process of serious games authoring and validation.

Serious games have proven several advantages when used in education improving students learning. However, games are still complex to deploy in the class for average teachers. Many teachers still do not see games as a powerful tool to improve their teaching work. To this end, it is essential to humanize the game technology making the use of games more transparent to teachers in a way that they get the benefits and avoid most of the game deployment complexity. We consider that Game Learning Analytics is one of the keys to help teachers in the application of serious games in the classrooms. Game Learning Analytics allows to capture data from students' interactions with games and derive information that simplify teachers' tasks. Doing it in a transparent way within the game environment (i.e. stealth assessment) can provide evidence-based data about the learners' knowledge at each point of time. Combining both game learning analytics in near real-time and offline, and stealth assessment for games, it could be possible to leverage their use in classroom settings at real-time making their use easier for teachers.

Teachers often wish to integrate activities from disparate sources into their courses. For example, gamified activities, mediated through technology, can promote the type of active learning required to develop higher-level engagement by students. However, unless the activities have been designed to facilitate it, integrating their analytics into a single dashboard can require significant development effort. A general solution to such heterogeneous analytics integration can be of great value, by presenting a single view of student actions throughout the different parts of a course. We describe the problems presented when integrating the analytics of three heterogeneous stand-alone activities, in the context of a EU project to improve software engineering teaching. The idea is to increase student engagement via gamification, and explore the design space of possible solutions for providing integrated analytics over the heterogeneous activities. We then describe the design of a proof-of-concept implementation, based on the use of both xAPI trackers and simple CSV files for information exchange, single sign-on, a minimal class management web application, and updates to the analytics platform to allow dynamic changes in the multi-level analysis. The resulting approach can be readily applied to similar heterogeneous scenarios.

Serious games validation is a highly complex and burdensome process. To ensure that games meet their intended educational goals, it is necessary to have a clear experimental design and the necessary tools to minimize the errors that may appear in the process. In this article, after describing the most common problems that we have found while validating our own games, we present Simva, a tool designed to simplify the process of validating serious games with formal questionnaires and relating them with learning analytics data, reducing time, cost, and error rates.

Game engines and developer-friendly authoring tools have greatly simplified entertainment game development. However, this does not extend to serious games which, in particular, require involvement of non-developer stakeholders. We present the first evaluation of uAdventure, an easy-to-use game development environment for narrative point-and-click graphic adventure games. uAdventure is a re-implementation of the previously-validated eAdventure environment on top of the Unity game engine. The idea is to get most of the advantages of the Unity professional environment at a fraction of its complexity, and without requiring programming knowledge to use it. uAdventure include educational-oriented affordances, such as assessment and learning analytics; and has been formatively evaluated by heterogeneous users with different degrees of technical knowledge. The results of our evaluation show much simpler story creation for profiles that had no previous knowledge of the engine, and positive feedback from more technical profiles which would use the tool as a prototyping tool for complex projects.

Game learning analytics has a great potential to provide insight and improve the use of games in different educational situations. However, it is necessary to clearly establish what the learner's requirements are and to set realistic expectations about the learning process and outcomes. Application of game learning analytics requires pedagogically informed policies that settle the learning goals and relate them to analysis and visualization; and a supporting infrastructure that provides the mechanism on top of which it is executed. Both concerns can be addressed separated: on the one hand, there is a Learning Analytics Model (LAM) which describes how the analysis is carried out, interpreted as learning, and presented to stakeholders; and on the other hand, an underlying analytics system that concentrates on performance, security, flexibility and generality. An important advantage of this separation is that it allows LAM authors to concentrate on their area of expertise, limiting their exposition to the actual mechanism used underneath. However, LAMs built for a single game fail to account for the frequent case where games and their analytics are aggregated into larger, overarching plots, games or courses. This work describes an extension to an existing game learning analytics system, used in RAGE and BEACONING H2020 projects, which manages multilevel analytics through improvements to both policy and mechanism; and introduces meta-Learning Analytic Models, which characterize learning in hierarchical structures.

Pervasive gaming has experimented a huge commercial growth with location-based game successes such as Pokémon GO or Ingress. The serious game industry has an opportunity to take advantage of location-based mechanics to better connect games with the real world, creating more authentic immersive learning environments. Games such as historical tours, story-based exploration, laboratories, or flora explorations can greatly benefit from location-based mechanics. Location-based games usually include an augmented map to provide game context, which combines both traditional game-dependent elements such as avatars, and location-based elements such as areas or points of interest. For players, interacting with location-based elements may involve entering or exiting specific areas; or reaching a certain location and looking in a specific direction. To include standards-based learning analytics for location-based serious games (SGs), we have added support for player movement and location-based interactions to the pre-existing xAPI serious game profile. We have validated this approach through a case-study example that guided players through different sports-related facilities within a large outdoor area. This work have been carried out and it is available as part of the analytics infrastructure used in EU H2020 RAGE and BEACONING serious game projects.

Applying games in education provides multiple benefits clearly visible in entertainment games: their engaging, goal-oriented nature encourages students to improve while they play. Educational games, also known as Serious Games (SGs) are video games designed with a main purpose other than pure entertainment; their main purpose may be to teach, to change an attitude or behavior, or to create awareness of a certain issue. As educators and game developers, the validity and effectiveness of these games towards their defined educational purposes needs to be both measurable and measured. Fortunately, the highly interactive nature of games makes the application of Learning Analytics (LA) perfect to capture students' interaction data with the purpose of better understanding or improving the learning process. However, there is a lack of widely adopted standards to communicate information between games and their tracking modules. Game Learning Analytics (GLA) combines the educational goals of LA with technologies that are commonplace in Game Analytics (GA), and also suffers from a lack of standards adoption that would facilitate its use across different SGs. In this paper, we describe two key steps towards the systematization of GLA: 1), the use of a newly-proposed standard tracking model to exchange information between the SG and the analytics platform, allowing reusable tracker components to be developed for each game engine or development platform; and 2), the use of standardized analysis and visualization assets to provide general but useful information for any SG that sends its data in the aforementioned format. These analysis and visualizations can be further customized and adapted for particular games when needed. We examine the use of this complete standard model in the GLA system currently under development for use in two EU H2020 SG projects.

Educational games (aka serious games, SG) are powerful educational contents. However, they are costly to develop, and once developed, SGs become dependent on software and hardware combinations that may become obsolete, such as Adobe Flash or Java Applets. Addressing these problems would allow a much greater use of SGs in education. The eAdventure authoring tool, developed by the e-UCM research group, addressed high development costs, and resulted in the creation of multiple SGs in collaboration with different institutions. However, eAdventure's Java Applets have become increasingly difficult to run due to platform obsolescence. To maintain the benefits of the eAdventure platform and user base, we have created new platform called uAdventure: an SG editor built on top of the game engine Unity that allows for the creation of educational adventure games without requiring programming. Since Unity is supported on a majority of platforms (including mobile). By developing SGs with uAdventure, the games become future-proof, as they can be updated and retargeted for new platforms as required. In this sense, uAdventure improves the lifecycle of SGs by reducing both authoring and maintenance costs.

Serious Games can benefit from the commercial video games industry by taking advantage of current development tools. However, the economics and requirements of serious games and commercial games are very different. In this paper, we describe the factors that impact the total cost of ownership of serious games used in educational settings, review the specific requirements of games used as learning material, and analyze the different development tools available in the industry highlighting their advantages and disadvantages that must be taken into account when using them to develop a serious game.

MOOCs are disrupting educational technology by democratizing access to high quality courses offered by some of the most prestigious universities. However, due to low entry cost and multiple other factors, MOOCs suffer from low retention rates: making MOOCs more interactive and engaging is therefore an important goal. Educational games provide environments that mix both immersion and quick feedback cycles, and we consider that games also have the potential to improve interaction and assessment in MOOCs environments; indeed, many game-like simulations have already been successfully applied within MOOCs in domains such as electronic circuits (schematic circuit simulator) and biology (foldit, eyewire). In this work, we analyze multiple such experiences and propose a catalogue of best practices that could contribute to the successful integration of educational games into MOOC platforms.

Massive Online Open Courses are in the process of radically changing the use of technology in education, as measured both by the number of enrolled students and the level of academic excellence of the institutions that are driving their adoption. Typically, online learning management systems have relied on simple questionnaires for feedback and evaluation. On the other hand, MOOC platforms such as EDX have, from the onset, provided a great flexibility in this regard, providing peer grading and several examples of highly interactive activities such as hands-on simulations. This paper explores the integration of serious games as a new type of MOOC activity, providing increased engagement and a valuable source of learning analytics. The inclusion of serious games has implications for both courses and games. Due to the diversity of existing serious game and MOOC platforms, we focus on the specific case of integrating EADVENTURE games into EDX.

This paper presents our perspective on serious games, covering relevant aspects of the current situation and what we consider are the key elements to be developed over the next few years. Serious games have undergone a long journey from niche research to application in real settings, coming closer to the generalization of games in mainstream education. We describe some of the lines of research we consider more relevant for completing this journey.

The Learning Analytics (LA) discipline analyzes educational data obtained from student interaction with online resources. Most of the data is collected from Learning Management Systems deployed at established educational institutions. In addition, other learning platforms, most notably Massive Open Online Courses such as Udacity and Coursera or other educational initiatives such as Khan Academy, generate large amounts of data. However, there is no generally agreedupon data model for student interactions. Thus, analysis tools must be tailored to each system's particular data structure, reducing their interoperability and increasing development costs. Some e-Learning standards designed for content interoperability include data models for gathering student performance information. In this paper, we describe how well-known LA tools collect data, which we link to how two e-Learning standards - IEEE Standard for Learning Technology and Experience API - define their data models. From this analysis, we identify the advantages of using these e-Learning standards from the point of view of Learning Analytics.

Visual overviews of tables of numerical and categorical data have been proposed for tables with a single value per cell. This article addresses the problem of exploring tables with columns that consist of cells that are distributions, for example, the distributions of movie ratings or trust ratings in recommender systems, age distributions in demographic data, usage distributions in logs of telephone calls, and so on. This article expands on heatmap approaches and proposes a novel way of displaying and interacting with distribution data. The usability study demonstrates the benefits of the heatmap interface in providing an overview of the data and facilitating the discovery of interesting clusters, patterns, outliers and relationships between columns.

Traditional network analysis tools support analysts in studying a single network. ManyNets offers these analysts a powerful new approach that enables them to work on multiple networks simultaneously. Several thousand networks can be presented as rows in a tabular visualization, and then inspected, sorted and filtered according to their attributes. The networks to be displayed can be obtained by subdivision of larger networks. Examples of meaningful subdivisions used by analysts include ego networks, community extraction, and time-based slices. Cell visualizations and interactive column overviews allow analysts to assess the distribution of attributes within particular sets of networks. Details, such as traditional node-link diagrams, are available on demand. We describe a case study analyzing a social network geared towards film recommendations by means of decomposition. A small usability study provides feedback on the use of the interface on a set of tasks issued from the case study.

Programming assignments are easy to plagiarize in such a way as to foil casual reading by graders. Graders can resort to automatic plagiarism detection systems, which can generate a matrix that covers all possible pairings. Most plagiarism detection programs then present this information as a simple ranked list, losing valuable information in the process.The Ac system uses the whole distance matrix to provide graders with multiple linked visualizations. The graph representation can be used to explore clusters of highly related submissions at different filtering levels. The histogram representation presents compact individual histograms for each submission, complementing the graph representation in aiding graders during analysis.Although Ac's visualizations were developed with plagiarism detection in mind, they should also prove effective to visualize distance matrices from other domains, as demonstrated by preliminary experiments.

Graphs provide good representations for many domains. Interactive graph-based interfaces are desireable to browse and edit data for these domains. However, as graphs increase in size, interactive interfaces risk information overload and low responsiveness. Focus+context approaches overcome these problems by presenting abridged views of the graph. Users can then navigate among views with a level-of-detail mechanism. If jumps from each view to the next are easy to follow, users will gain a good mental map of the whole graph; otherwise, they may become disoriented.In this work, we identify three factors that affect mental map preservation during navigation of interactive focus+context graphs: the predictability of navigational actions, the degree of change from one view to the next, and the traceability of changes once they occur. Strategies for preserving user orientation are classified according to these factors, and new strategies developed for the CLOVER visualization environment are presented.

In this paper, we present an automatic procedure to generate students’ knowledge conceptual models from their answers to an automatic free-text scoring system. The conceptual model is defined as a simplified representation of the concepts and relationships among them that each student keeps in his or her mind about an area of knowledge. It is considered that each area of knowledge comprises several topics and each topic several concepts. Each concept can be identified by a term that the students should use. A concept can belong to one topic or to several topics. The conceptual model is graphically displayed to the teachers as a conceptual map so that they can instantly see which concepts have already been assimilated and which ones should still be reviewed as they have been misunderstood.

Complex hypermedia structures can be difficult to author and maintain, especially when the usual hierarchic representation cannot capture important relations. We propose a graph-based direct manipulation interface that uses multiple focus+context techniques to avoid display clutter and information overload. A semantical fisheye lens based on hierarchical clustering allows the user to work on high-level abstracts of the structure. Navigation through the resulting graph is animated in order to avoid loss of orientation, with a force-directed algorithm in charge of generating successive layouts. Multiple views can be generated over the same data, each with independent settings for filtering, clustering and degree of zoom.While these techniques are all well-known in the literature, it is their combination and application to the field of hypermedia authoring that constitutes a powerful tool for the development of next-generation hyperspaces.A generic framework, CLOVER, and two specific applications for existing hypermedia systems have been implemented.