Active mobility, such as biking, faces a common challenge in Swedish municipalities due to the lack of adequate lighting during the dark winter months. Insufficient lighting infrastructure hinders individuals from choosing bicycles, despite the presence of well-maintained bike paths and a willingness to cycle. To address this issue, a project has been undertaken in the Swedish municipality of Skara for an alternative lighting solution using drones. A series of tests have been conducted based on drone prototypes developed for the selected bike paths. Participants were invited to cycle in darkness illuminated by drone lighting and share their mobility preferences and perception. This paper summarizes the users’ perception of drone lighting as an alternative to fixed lighting on bike paths, with a special focus on the impact on travel habits and the perceived sense of security and comfort. Most participants were regular cyclists who cited bad weather, time, and darkness as significant factors that deterred them from using bicycles more frequently, reducing their sense of security. With drone lighting, the participants appreciated the illumination’s moonlight-like quality and its ability to enhance their sense of security by illuminating the surroundings. On the technology side, they gave feedback on reducing the drone’s sound and addressing lighting stability issues. In summary, the test results showcase the potential of drone lighting as a viable alternative to traditional fixed lighting infrastructure, offering improved traffic safety, sense of security, and comfort. The results show the feasibility and effectiveness of this innovative approach, supporting transformation towards active and sustainable mobility, particularly in regions facing lighting challenges.

The workshop features presentations of accepted contributions to the RecSys Challenge 2021, organized by Politecnico di Bari, ETH Zürich, Jönköping University, and the data set is provided by Twitter. The challenge focuses on a real-world task of tweet engagement prediction in a dynamic environment. For 2021, the challenge considers four different engagement types: Likes, Retweet, Quote, and replies. This year's challenge brings the problem even closer to Twitter's real recommender systems by introducing latency constraints. We also increases the data size to encourage novel methods. Also, the data density is increased in terms of the graph where users are considered to be nodes and interactions as edges. The goal is twofold: to predict the probability of different engagement types of a target user for a set of Tweets based on heterogeneous input data while providing fair recommendations. In fact, multi-goal optimization considering accuracy and fairness is particularly challenging. However, we believed that the recommendation community was nowadays mature enough to face the challenge of providing accurate and, at the same time, fair recommendations. To this end, Twitter has released a public dataset of close to 1 billion data points, > 40 million each day over 28 days. Week 1-3 will be used for training and week 4 for evaluation and testing. Each datapoint contains the tweet along with engagement features, user features, and tweet features. A peculiarity of this challenge is related to keeping the dataset updated with the platform: if a user deletes a Tweet, or their data from Twitter, the dataset is promptly updated. Moreover, each change in the dataset implied new evaluations of all submissions and the update of the leaderboard metrics. The challenge was well received with 578 registered users, and 386 submissions.

Various social influences affect group decision-making processes. For instance, individuals may adapt their behavior to fit in with the group's majority opinion. Furthermore, ingroup favoritism may lead individuals to favor the ideas of ingroup members rather than the outgroup. So far, little is explored on how these phenomena of social conformity and ingroup favoritism manifest in group decision-making processes when a group has to decide in favor or against an item. We address such a scenario where the 'flipping direction' of conformity (in favor or against an item) matters. Specifically, we explore whether and how the ingroup favoritism manifests differently in terms of conformity behavior depending on the 'flipping direction'. The results show that group inclusiveness does not play a role in the general tendency to conform. However, when it comes to a negative flipping direction, a higher feeling of group inclusiveness seems to play a role; yet, for individualist cultures only.

A strong research record on conformity has evidenced that individuals tend to conform with a group’s majority opinion. In contrast to existing literature that investigates conformity to a majority group opinion against an objectively correct answer, the originality of our study lies in that we investigate conformity in a subjective context. The emphasis of our analysis lies on the concept of “switching direction” in favor or against an item. We present first results from an online experiment where groups of five had to create a music playlist. A song was added to the playlist with an unanimous positive decision only. After seeing the other group members’ ratings, participants had the opportunity to revise their own response. Our results suggest different conformity behaviors for originally favored compared to disliked songs. For favored songs, one negative judgement by another group member was sufficient to induce participants to downvote the song. For originally disliked songs, in contrast, a majority of positive judgements was needed to induce participants to switch their vote.

After the success the RecSys 2020 Challenge, we are describing anovel and bigger dataset that was released in conjunction with theACM RecSys Challenge 2021. This year’s dataset is not only bigger(~1B data points, a 5 fold increase), but for the first time it take intoconsideration fairness aspects of the challenge. Unlike many staticdatsets, a lot of effort went into making sure that the dataset wassynced with the Twitter platform: if a user deleted their content,the same content would be promptly removed from the dataset too.In this paper, we introduce the dataset and challenge, highlightingsome of the issues that arise when creating recommender systemsat Twitter scale. 

Business intelligence (BI) systems are software applications that are used to gather and process data and to deliver the processed data in understandable way to the end users. With a younger generation of users moving into key positions in organizations and enterprises higher user experience (UX) demands are placed on BI systems interfaces. Companies developing BI systems lack standardized routines for implementing UX in their BI solutions. The purpose of this study was to develop a theoretical framework based on existing research and combine it with empirical data gathered from professionals in BI systems industry in Sweden with the intention of proposing a UX framework applicable to BI systems development. The study resulted in a framework being developed using iterative build-evaluate iterations. The framework is a scalable UX framework for BI systems interfaces covering areas from planning and strategizing to implementation, maintenance, and evaluation.

In group decision-making, we can frequently observe that an individual adapts their behavior or belief to fit in with the group’s majority opinion. This phenomenon has been widely observed to exist especially against an objectively correct answer—in face-to-face and online interaction alike. To a lesser extent, studies have investigated the conformity effect in settings based on personal opinions and feelings; thus, in settings where an objectively right or wrong answer does not exist. In such settings, the direction of conformity tends to play a role in whether an individual will conform. While cultural differences in conformity behavior have been observed repeatedly in settings with an objectively correct answer, the role of culture has not been explored yet for settings with subjective topics. Hence, the focus of this study is on how conformity develops across cultures for such cases. We developed an online experiment in which participants needed to reach a positive group consensus on adding a song to a music playlist. After seeing the group members’ ratings, the participants had the opportunity to revise their own. Our findings suggest that the willingness to flip to a positive outcome was far less than to a negative outcome. Overall, conformity behavior was far less pronounced for participants from the United Kingdom compared to participants from India.

With the vast amount of online content available to us, platforms are utilizing recommender systems to help their users in their decision making. By presenting content that is in line with the user’s taste and preferences a personalized experience can be created. Platforms such as Netflix and Spotify have started to create an additional layer of personalization by framing of the content presentation through tailoring album art and titles towards the user. Even though all content in a user account is personalized, this additional layer of personalization creates a distinction between “regular" content and implied personalized content. In this work we explore how the textual framing (generic vs. personalized) of music playlists influences behaviors and content expectations. Our findings show that users mostly ignore the implied personalized playlists as they expect that it only consists of previously listened songs, while generic playlists are exploited to find new music to listen to.

Psychological models are increasingly being used to explain online behavioral traces. Aside from the commonly used personality traits as a general user model, more domain dependent models are gaining attention. The use of domain dependent psychological models allows for more fine-grained identification of behaviors and provide a deeper understanding behind the occurrence of those behaviors. Understanding behaviors based on psychological models can provide an advantage over data-driven approaches. For example, relying on psychological models allow for ways to personalize when data is scarce. In this preliminary work we look at the relation between users' musical sophistication and their online music listening behaviors and to what extent we can successfully predict musical sophistication. An analysis of data from a study with 61 participants shows that listening behaviors can successfully be used to infer users' musical sophistication.

Information retrieval and recommender systems are deployed in real world environments. Therefore, to get a real feeling for the system, we should study their characteristics in “real world studies”. This raises the question: What does it mean for a study to be realistic? Does it mean the user has to be a real user of the system or can anyone participate in a study of the system? Does it mean the system needs to be perceived as realistic by the user? Does it mean the manipulations need to be perceived as realistic by the user?

Recommender systems are designed to help us navigate through an abundance of online content. Collaborative filtering (CF) approaches are commonly used to leverage behaviors of others with a similar taste to make predictions for the target user. However, CF is prone to introduce or amplify popularity bias in which popular (often consumed or highly ranked) items are prioritized over less popular items. Many computational metrics of popularity biases — and resulting algorithmic (un)fairness — have been presented. However, it is largely unclear whether these metrics reflect human perception of bias and fairness. We conducted a user study with 170 participants to explore how users perceive recommendation lists created by algorithms with different degrees of popularity bias. Our results show — surprisingly — that popularity biases in recommendation lists are barely observed by users, even when corresponding bias/fairness metrics clearly indicate them. 

Driving monitoring and assistance systems are increasingly implemented by car manufacturers to increase safety and comfort for car drivers. Through notifications such systems support or create awareness in different driving situations. To provide appropriate notifications, knowledge about the driver’s needs need to be gained. In this study we investigate the acceptance of certain notifications in several driving scenarios for different driving styles. Through focus groups we found that there are different notification needs based on driving styles in relation to different driving scenarios. However, our results suggest that notification needs are more influenced by the cognitive load that is used rather than driving style on its own. Furthermore, deeper knowledge should be gained on the negative effects of providing notifications as there are situations in which a driver is rather be left alone than being assisted through notifications.

With life expectancy steadily increasing, healthy aging is becoming more important. Especially at a later age, the susceptibility to complications, such as morbidity, increases. Engaging in sufficient physical activities throughout a lifespan lowers the chances on such complications and contributes to an increased quality of life. However, the vast majority of the world's population does not engage enough in any form of physical activity. In this position paper we propose a serious game solution to promote physical activities based on the popular Tamagotchi from the 90's. We propose the virtual character of which the user needs to take care of to be a reflection of oneself. Thereby, any (in)activities of the user is directly reflected through the emotional and physical state of the character. Through the character, we hope to increase engagement in physical activities and facilitate long term behavioral change. Furthermore, we propose additional features and open research questions.

Due to the rise of available online music, a lot of music consumption is moving from traditional offline media to online sources. Online music sources offer almost an unlimited music collection to its users. Hence, how music is consumed by users (e.g., experts) may differ from traditional offline sources. In this work we explored how musically sophisticated users (i.e. experts) consume online music in terms of diversity. To analyze this, we gathered data from two different sources: Last.fm and Spotify. As expertise is defined by the ubiquitousness of experiences, we calculated different diversity measurements to explore how ubiquitous (in terms of diversity) the listening behaviors of users are. We found that different musical sophistication levels correspond to applying diversity related to specific kind of musical characteristics (i.e., artist or genre). Our results can provide knowledge on how systems should be designed to provide better support to expert users.

Instagram is a popular social networking application that allows users to express themselves through the uploaded content and the different filters they can apply. In this study we look at personality prediction from Instagram picture features. We explore two different features that can be extracted from pictures: 1) visual features (e.g., hue, valence, saturation), and 2) content features (i.e., the content of the pictures). To collect data, we conducted an online survey where we asked participants to fill in a personality questionnaire and grant us access to their Instagram account through the Instagram API. We gathered 54,962 pictures of 193 Instagram users. With our results we show that visual and content features can be used to predict personality from and perform in general equally well. Combining the two however does not result in an increased predictive power. Seemingly, they are not adding more value than they already consist of independently.

Instagram is a popular social networking application that allowsusers to express themselves through the uploaded contentand the different filters they can apply. In this study we look atthe relationship between the content of the uploaded Instagrampictures and the personality traits of users. To collect data, weconducted an online survey where we asked participants tofill in a personality questionnaire, and grant us access to theirInstagram account through the Instagram API. We gathered54,962 pictures of 193 Instagram users. Through the GoogleVision API, we analyzed the pictures on their content and clusteredthe returned labels with the k-means clustering approach.With a total of 17 clusters, we analyzed the relationship withusers’ personality traits. Our findings suggest a relationshipbetween personality traits and picture content. This allow fornew ways to extract personality traits from social media trails,and new ways to facilitate personalized systems.

Personality traits are increasingly being incorporated in systems to provide a personalized experience to the user. Current work focusing on identifying the relationship between personality and behavior, preferences, and needs often do not take into account differences between age groups. With music playing an important role in our lives, differences between age groups may be especially prevalent. In this work we investigate whether differences exist in music listening behavior between age groups. We analyzed a dataset with the music listening histories and personality information of 1415 users. Our results show agreements with prior work that identi€ed personality-based music listening preferences. However, our results show that the agreements we found are in some cases divided over different age groups, whereas in other cases additional correlations were found within age groups. With our results personality-based systems can provide better music recommendations that is in line with the user’s age.

The number of people that have been in touch with drugs is continuously increasing. Excessive intake of drugs becomes problematic when it turns into disorderly behaviors, such as addictions. In order to treat these disorderly behaviors, treatment plans often adhere to a one-size-fits-all approach with fixed and standardized steps. However, for effective treatment of disorderly behaviors it has been acknowledged that personalized treatment programs are necessary. The personality of people has been argued to be a factor that plays an important role in setting up effective treatment plans. In this work we explored the predictability of people’s personality traits based on their drug consumption profile. Based on self-reported consumption frequencies of "abusable psychoactive drugs," we found among 1878 respondents that drug consumption profiles can be used to predict people’s personality traits. The prediction of personality traits can be used to circumvent intruding questionnaires and to implicitly create personalized treatment programs.

The sixth HUMANIZE workshop1 on Transparency and Explainability in Adaptive Systems through User Modeling Grounded in Psychological Theory took place in conjunction with the 27th annual meeting of the Intelligent User Interfaces (IUI)2 community that was hosted virtually by the University of Helsinki (Finland) on March 22, 2022. The 2022 edition of the workshop was held together with TExSS (Transparency and Explanations in Smart Systems) 3. The workshop provided a venue for researchers from different fields to interact by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of two papers was accepted for this edition of the workshop.

Music streaming services increasingly incorporate different ways for users to browse for music. Next to the commonly used “genre” taxonomy, nowadays additional taxonomies, such as mood and activities, are often used. As additional taxonomies have shown to be able to distract the user in their search, we looked at how to predict taxonomy preferences in order to counteract this. Additionally, we looked at how the number of categories presented within a taxonomy influences the user experience. We conducted an online user study where participants interacted with an application called “Tune-A-Find”. We measured taxonomy choice (i.e., mood, activity, or genre), individual differences (e.g., personality traits and music expertise factors), and different user experience factors (i.e., choice difficulty and satisfaction, perceived system usefulness and quality) when presenting either 6- or 24-categories within the picked taxonomy. Among 297 participants, we found that personality traits are related to music taxonomy preferences. Furthermore, our findings show that the number of categories within a taxonomy influences the user experience in different ways and is moderated by music expertise. Our findings can support personalized user interfaces in music streaming services. By knowing the user’s personality and expertise, the user interface can adapt to the user’s preferred way of music browsing and thereby mitigate the problems that music listeners are facing while finding their way through the abundance of music choices online nowadays.

Nowadays, the profound digital transformation has upgraded the role of the computational system into an intelligent multidimensional communication medium that creates new opportunities, competencies, models and processes. The need for human-centered adaptation and personalization is even more recognizable since it can offer hybrid solutions that could adequately support the rising multi-purpose goals, needs, requirements, activities and interactions of users. The HAAPIE workshop1 embraces the essence of the "human-machine co-existence"and brings together researchers and practitioners from different disciplines to present and discuss a wide spectrum of related challenges, approaches and solutions. In this respect, the seventh edition of HAAPIE includes 5 long papers and 2 short papers.

Nowadays, the profound digital transformation has upgraded the role of the computational system into an intelligent multidimensional communication medium that creates new opportunities, competencies, models and processes. The need for human-centered adaptation and personalization is even more recognizable since it can offer hybrid solutions that could adequately support the rising multi-purpose goals, needs, requirements, activities and interactions of users. The HAAPIE workshop1 embraces the essence of the "human-machine co-existence"and brings together researchers and practitioners from different disciplines to present and discuss a wide spectrum of related challenges, approaches and solutions. In this respect, the seventh edition of HAAPIE includes 2 long papers and 5 short papers.

Nowadays, the profound digital transformation has upgraded the role of the computational system into an intelligent multidimensional communication medium that creates new opportunities, competencies, models and processes. The need for human-centered adaptation and personalization is even more recognizable since it can offer hybrid solutions that could adequately support the rising multi-purpose goals, needs, requirements, activities and interactions of users. The HAAPIE workshop1 embraces the essence of the "human-machine co-existence"and brings together researchers and practitioners from different disciplines to present and discuss a wide spectrum of related challenges, approaches and solutions. In this respect, the ninth edition of HAAPIE includes 4 long papers and 2 short papers.

This study investigated the global changes in online music listening behaviors in response to COVID-19 and its restrictions (such as quarantine, school and workplace closures, and travel restrictions). In addition, the research included an examination of how friendship networks and online communication motives have moderated the effect of COVID-19 on music listening behaviors. The causal inference methods: difference in differences (DiD) and two-way fixed effects (TWFE), were conducted to analyze the online music listening behaviors and social interactions of 37,328 Last.fm users in 45 countries before and after the first wave of confinement. It was found that in response to COVID-19, the quantity, variety, and novelty of music consumption decreased, shifting toward mainstream artists, whereas individuals with more online social connections and communications showed the reverse behavior. Our research shows that online social interactions and community development significantly impact listeners’ behaviors and can be used as a guide to developing new design strategies for digital media, such as music, movies, and games. 

The success of online music platforms depends on the strength of the recommendation systems (RSs) that employ users’ interaction data to offer customised music listening experiences. Traditional recommendation systems, however, assume users are independent actors and identically distributed, ignoring social interactions or connections among users and the role of the Network Effect (NE). In this study, leveraging the social network theory and utilising the attributes of an online music platform, we investigate the impact of network effect on the patterns of music listening. As a result, we propose a new approach for measuring the network effect (Universal Network Effect) as a function of the structure of the network, data-driven learning, and improvements realised with Artificial Intelligence (AI), to scrutinize network effect at both individual and whole network level. The everevolving approach to network effect measurement enables us to further the study to determine the changes in user behaviour.

The fifth HUMANIZE workshop1 on Transparency and Explainability in Adaptive Systems through User Modeling Grounded in Psychological Theory took place in conjunction with the 26th annual meeting of the Intelligent User Interfaces (IUI)2 community in Texas, USA on April 17, 2021. The workshop provided a venue for researchers from different fields to interact by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of five papers was accepted for this edition of the workshop.

The fourth HUMANIZE workshop1 on Transparency and Explainability in Adaptive Systems through User Modeling Grounded in Psychological Theory took place in conjunction with the 25th annual meeting of the Intelligent User Interfaces (IUI)2 community in Cagliari, Italy on March 17, 2020. The workshop provided a venue for researchers from different fields to interact by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of four papers was accepted for this edition of the workshop. 

Personalized systems are systems that adapt themselves to meet the inferred needs of individual users. The majority of personalized systems mainly rely on data describing how users interacted with these systems. A common approach is to use historical data to predict users’ future needs, preferences, and behavior to subsequently adapt the system to cater to these predictions. However, this adaptation is often done without leveraging the theoretical understanding between behavior and user traits that can be used to characterize individual users or the relationship between user traits and needs that can be used to adapt the system. Adopting a more theoretical perspective can benefit personalization in two ways: (i) letting systems rely on theory can reduce the need for extensive data-driven analysis, and (ii) interpreting the outcomes of data-driven analysis (such as predictive models) from a theoretical perspective can expand our knowledge about users. However, incorporating theoretical knowledge in personalization brings forth a number of challenges. In this chapter, we review literature that taps into aspects of (i) psychological models from traditional psychological theory that can be used in personalization, (ii) relationships between psychological models and online behavior, (iii) automated inference of psychological models from data, and (iv) how to incorporate psychological models in personalized systems. Finally, we propose a step-by-step approach on how to design personalized systems that take users’ traits into account.

The second workshop on Theory-Informed User Modeling for Tailoring and Personalizing Interfaces (HUMANIZE) took place in conjunction with the 23rd annual meeting of the intelligent user interfaces (IUI) community in Tokyo, Japan on March 11, 2018. The goal of the workshop was to attract researchers from different fields by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of eight papers were accepted for this edition of the workshop.

The second workshop on Theory-Informed UserModeling for Tailoring and Personalizing Interfaces (HUMANIZE)1 took place in conjunction with the 23rd annual meeting of the intelligent user interfaces (IUI)2 community in Tokyo, Japan on March 11, 2018. The goal of the workshop was to attract researchers from different fields by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of eight papers were accepted for this edition of the workshop. 

The third workshop on Theory-Informed User Modeling for Tailoring and Personalizing Interfaces (HUMANIZE)1 took place in conjunction with the 24th annual meeting of the intelligent user interfaces (IUI)2 community in Los Angeles, CA, USA on March 20, 2019. The goal of the workshop was to attract researchers from different fields by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of six papers were accepted for this edition of the workshop.

Research has demonstrated that different types of users have different needs when it comes to personalized systems. Factors that have been argued to play a role are user traits and characteristics, such as personality and domain expertise. Within this work we explored the influence of listeners’ active engagement with music, or the extent to which people invest in enjoying music, on the relationship between item popularity and satisfaction with lists of recommendation. Using an online survey built on top of the Spotify API functionality we gathered participants’ most listened tracks and used those to compile playlists containing either popular or non-popular tracks. Our results show that active engagement plays a moderating role on the relationship between popularity and satisfaction, which can more specifically be explained by the extent to which popular songs allow people to discover their musical taste. Where listeners with low active engagement are limited in their discovery by tracks they are familiar with, those with high active engagement can actually use music they are familiar with to further discover their taste. Hence, for low active engagement listeners the most attractive recommendation lists are lists that strike a balance between familiar items and items that enable people to refine their musical taste.

Personalized systems are systems that adapt themselves to meet the inferred needs of individual users. The majority of personalized systems mainly rely on data describing how users interacted with these systems. A common approach is to use historical data to predict users’ future needs, preferences and behavior to subsequently adapt the system to cater to these predictions. However, this adaptation is often done without leveraging the theoretical understanding between behavior and user traits that can be used to characterize individual users or the relationship between user traits and needs that can be used to adapt the system. Adopting a more theoretical perspective can benefit personalization in three ways: (i) relying on theory can reduce the amount of data required to train compared to a purely data-driven system, (ii) interpreting the outcomes of data-driven analysis (such as predictive models) from a theoretical perspective can expand our knowledge about users and (iii) provide means for explanations and transparency. However, in order to incorporate theoretical knowledge in personalization a number of obstacles need to be faced. In this chapter, we review literature that taps into aspects of (i) psychological models from traditional psychological theory that can be used in personalization, (ii) relationships between psychological models and online behavior, (iii) automated inference of psychological models from data, and (iv) how to incorporate psychological models in personalized systems. Finally, we propose a step-by-step approach on how to design personalized systems that take users’ traits into account.

The third workshop on Theory-Informed User Modeling for Tailoring and Personalizing Interfaces (HUMANIZE) 1 took place in conjunction with the 24th annual meeting of the intelligent user interfaces (IUI) 2 community in Los Angeles, CA, USA on March 20, 2019. The goal of the workshop was to attract researchers from different fields by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of six papers were accepted for this edition of the workshop. 

In recent years, the integration of physical and cyber spaces has rapidly progressed. Sensing data from various sources is increasingly being fed back into physical space, and there are increasing cases of sensing and utilizing human biosignal information and emotions as well. This study investigates the psychological effects of visualizing and sharing people’s emotions and heart rate through two case studies. The first case study examines the psychological effects and usefulness of sharing others’ emotion and heart rate by displaying them beside he or she in physical space using a prototype of an augmented reality application called “Emo-Space”. The second case study proposes chat systems called “Emo-Circle” and “Heart-View” that allow for the visualization and sharing of emotions and heart rate and examines the effects of using the system on interpersonal relationships and self-awareness of emotions. We provide insights from the two case studies into that how such data of emotions and heart rate data indicating the physical and mental states of individuals can be applied to daily lives, including communication with others.

The development of products with a good user experience requires a thorough understanding of the prospective users' behaviors, preferences, and needs. One of the design approaches that places emphasis on the needs of users is the user-centered design process. However, there is a general resistance in organizations to incorporate user-centered design practices in product development. One influencing factor is that user-centered design is multidisciplinary. Hence, creates challenges to get a mutual understanding and collaboration across different stakeholders throughout the development. In this paper we present the results of a serious game prototype that was created to describe user-centered design across stakeholders. Results of the prototype evaluation reveal that it has potential to impart knowledge on concepts related to user-centered design. Additionally, we propose further development of the game by personalizing game elements to increase the effectiveness of learning and player experience.

Reproducibility of results is a central pillar of scientific work. In music information retrieval research, this is widely acknowledged and practiced by the communityby re-implementing algorithms and re-validating machine learning experiments. In this paper, we argue for an increased need to also reproduce the results and findings of user studies, including qualitative work, especially since these often lay the foundations and serve as justification for choices taken in algorithmic design and optimization criteria. As an example, we attempt to reproduce the study by Kim et al. [1] presented in the RecSys (2020) paper "Do Channels Matter? Illuminating Interpersonal Influence on Music Recommendations". By repeating this study on how interpersonal relationships can affect a user’s assessment of music recommendations on a new sample of n = 142 participants, we can largely confirm and support the validity of the original results. At the same time, we extend the analysis and also observe differences with regards to adoption rates between different channels as well as different factors that influences the adoption rate. From this specific reproducibility study, we conclude that potential cultural differences should be accounted for more explicitly in future studies and that systems development should be more explicitly connected to its intended target audience. 

Music recommender systems are a widely adopted application of personalized systems and interfaces.By tracking the listening activity of their users and building preference profiles, a user can be given recommendations based on the preference profiles of all users (collaborative filtering), characteristics of the music listened to (content-based methods), meta-data and relational data (knowledge-based methods; sometimes also considered content-based methods) or a mixture of these with other features (hybrid methods).In this chapter, we focus on the listener's aspects of music recommender systems.We discuss different factors influencing relevance for recommendation on both the listener's and the music's side and categorize existing work. In more detail, we then review aspects of (i) listener background in terms of individual, i.e., personality traits and demographic characteristics, and cultural features, i.e., societal and environmental characteristics, (ii) listener context, in particular modeling dynamic properties and situational listening behavior, and (iii) listener intention, in particular by studying music information behavior, i.e., how people seek, find, and use music information.This is followed by a discussion of user-centric evaluation strategies for music recommender systems. We conclude the chapter with a reflection on current barriers, by pointing out current and longer-term limitations of existing approaches and outlining strategies for overcoming these.

Music recommender systems are a widely adopted application of personalized systems and interfaces. By tracking the listening activity of their users and building preference profiles, a user can be given recommendations based on the preference profiles of all users (collaborative filtering), characteristics of the music listened to (contentbased methods), meta-data and relational data (knowledge-based methods; sometimes also considered content-based methods) or a mixture of these with other features (hybrid methods). In this chapter, we focus on the listener’s aspects of music recommender systems. We discuss different factors influencing relevance for recommendation on both the listener’s and the music’s side and categorize existing work. In more detail, we then review aspects of (i) listener background in terms of individual, i. e., personality traits and demographic characteristics, and cultural features, i. e., societal and environmental characteristics, (ii) listener context, in particular modeling dynamic properties and situational listening behavior and (iii) listener intention, in particular by studying music information behavior, i. e., how people seek, find and use music information. This is followed by a discussion of user-centric evaluation strategies for music recommender systems. We conclude the chapter with a reflection on current barriers, by pointing out current and longer-term limitations of existing approaches and outlining strategies for overcoming these.

The RecSys 2022 Challenge was a session-based recommendation task in the fashion domain. The dataset was supplied by Dressipi. Given session data consisting of views and purchases, as well as content data representing the fashion characteristics of the items, the task was to predict which item was purchased at the end of the session. The challenge ran for 3 months with a public leaderboard and final result on a separate hidden test set. There were over 300 teams that submitted a solution to the leaderboard and about 50 that submitted a solution for the final test set. The winning team achieved a MRR score of 0.216 which means that the correct target item was on average ranked 5th in the list of predictions. We identify some interesting common themes among the solutions in this paper and the winning approaches are presented in the workshop.

As part of the RecSys Challenge 2022, the Dressipi 1M Fashion Sessions dataset is publicly released. This paper gives an overview of the content and structure of the dataset, as well as explaining the process by which it was constructed. The dataset contains anonymous browsing sessions, a purchase for each session, as well as content data of the items. The content data consists of IDs that represent descriptive fashion characteristics of the items and have been assigned using Dressipi’s human-in-the-loop labelling system. We hope that this dataset will be valuable in recommender systems research beyond the RecSys Challenge and encourage more publications in the fashion domain.

With the development in technology and the increasing ubiquityof social media services, it has created new opportunities to studypersonality from the digital traces individuals leave behind. Thelarge number of user-generated images on social media hasprompted renewed interests in understanding the psychologicalfactors driving production and consumption behaviours of visualcontent. Instagram is currently the fastest growing photo-sharingsocial media platform, with more than 400 million active usersand nearly 100 million photos shared on the platform daily [1],and generates 1.2 billion likes each day [2]. The understanding ofthe appeal of visual content at an individual level is highlyrelevant to psychometric assessment, social media marketing andinterface personalisation. In this position paper, we address theneed to explore the avenue of automatic personality assessmentusing ‘liked’ images on Instagram.

Popularity bias in recommendation lists refers to over-representation of popular content and is a challenge for many recommendation algorithms. Previous research has suggested several offline metrics to quantify popularity bias, which commonly relate the popularity of items in users’ recommendation lists to the popularity of items in their interaction history. Discrepancies between these two factors are referred to as popularity miscalibration. While popularity metrics provide a straightforward and well-defined means to measure popularity bias, it is unknown whether they actually reflect users’ perception of popularity bias.

To address this research gap, we conduct a crowd-sourced user study on Prolific, involving 56 participants, to (1) investigate whether the level of perceived popularity miscalibration differs between common recommendation algorithms, (2) assess the correlation between perceived popularity miscalibration and its corresponding quantification according to a common offline metric. We conduct our study in a well-defined and important domain, namely music recommendation using the standardized LFM-2b dataset, and quantify popularity miscalibration of five recommendation algorithms by utilizing Jensen-Shannon distance (JSD). Challenging the findings of previous studies, we observe that users generally do perceive significant differences in terms of popularity bias between algorithms if this bias is framed as popularity miscalibration. In addition, JSD correlates moderately with users’ perception of popularity, but not with their perception of unpopularity.

In this paper, we describe the LFM-1b User Genre Profile dataset. It provides detailed information on musical genre preferences for more than 120,000 listeners and links to the LFM-1b dataset. We created the dataset by exploiting social tags, indexing them using two genre term sets, and aggregating the resulting annotated listening events on the user level. We foresee several applications of the dataset in music retrieval and recommendation tasks, among others to build and evaluate decent user models, to alleviate cold-start situations in music recommender systems, and to increase their performance using the additional abstraction layer of genre. We further present results of statistical analyses of the dataset, regarding genre preferences and their consistencies. We do so for the entire user population and for user groups defined by demographic similarities. Moreover, we report interesting insights about correlations between musical preferences on the genre level.

Considering the cultural background of users is known to improve recommender systems for multimedia items. In this work, we focus on music and analyze user demographics and music listening events in a large corpus (120,000 users, 10⁹ events) from Last.fm to investigate whether similarity between countries in terms of cultural and socio-economic factors is reflected in music taste. To this end, we propose a tag-based model to describe the music taste of a country and correlate the resulting music profiles to Hofstede’s cultural dimensions and the Quality of Government data. Spearman’s rank-order correlation and Quadratic Assignment Procedure indeed indicate statistically significant weak to medium correlations of music taste and several cultural and socio-economic factors. The results will help elaborating culture-aware models of music listeners and in turn likely yield improved music recommender systems.

Companies face several digital communication challenges when it comes to promoting green products or services. The framing effect, which refers to the presentation of information, can significantly influence decision-making in digital interfaces. This research explores the impact of information framing through text and visuals on purchase decisions for sustainable fashion products. An online evaluation study ( = 84) of an e-commerce environment was conducted. We found that visual framing significantly affected user product choices, supporting more sustainability decisions. In contrast, little evidence was found that supported the effectiveness of linguistic (i.e., message) framing on user product choices. We discuss implications on how product pages should be designed to encourage sustainable decision-making.

Recommender systems are prone to triggering choice overload among users due to the typically largeset sizes. Various applications have been developed that aim to overcome this through interface design,notably by so-called multi-list recommender systems. However, to what extent such user interface designactually reduces choice overload compared to single-list interfaces has yet to be examined. In a user study(𝑁 = 150), we compared three common user interfaces (UIs) in the context of recipe recommendation: asingle-list UI, a grid UI and a multi-list UI. Whereas earlier studies found differences in choice difficultyand choice satisfaction across grid-based and multi-list recommender interfaces, we observed no suchdifferences, as the explanations were possibly not sufficiently helpful. Instead, we found that grid-basedUIs and multi-list UIs had a higher perceived ease of use than a single-list UI, which in turn reducedchoice difficulty. The benefits of such interfaces, thus, may lie in the organization of the UI, at least inthe recipe domain. 

One of the important aspects of movie-making is to trigger emotional responses in viewers. These emotional experiences can be divided into hedonic and eudaimonic. While the former are characterized as plain enjoyment, the latter deal with getting greater insight, self-reflection or contemplation. So far, modeling of user preferences about movies and personalization algorithms have largely ignored the eudaimonic aspect of the consumption of movies. In this paper we fill this gap by exploring what are the relationship between (i) eudaimonic and hedonic characteristics of movies, (ii) users' preferences and (iii) users' personality. Our results show that eudaimonic user profiling effectively divides users into pleasure-seekers and meaning-seekers.

Most of the research in recommender systems focuses on data-driven approaches. In this paper we present our vision for complementing data-driven approaches with model-driven ones. We present a preliminary experimental set-up and we expose our research plan. In the experimental set-up we acquired eudaimonic characteristics of movies and user preferences. Furthermore, we performed a preliminary analysis of the acquired data.