Urban spaces and urban lives are being transformed through “datafication,” which drives the design of data interaction tools and shapes how we make sense of the present and image our futures. Dominant narratives of data science influence the design of these tools, framing data as objective, neutral, and solutions to disentangle complex problems. Critical design and data researchers challenge these ideas and argue how such design ideologies create injustice and marginalize other ways of knowing. Through three interrelated projects, I investigate ways of designing data interactions and visualizations experiences that lean into creative practices to challenge the framing of objectivity, neutrality, and open up design space to foreground data and design as an entanglement of emotion, power and politics. The projects contribute to HCI by offering empirical insights into how people make sense of and use data, and by examining the role of creative practices in shaping data engagement.

The growth of AI does not come without cost. While we hear about the economic costs, the environmental and social costs are often obfuscated by mainstream narratives, and even when acknowledged, are accompanied by a sense of helplessness. To counter this, we, a group of designers and researchers, reflect on our experience leading a humour-based creative campaign surfacing the material impact of AI infrastructures. Through graphics design, physical installations, digital content creation and co-creation workshops, we leaned into humour as a creative practice to provoke collective reflection. Drawing on event ethnography, surveys and interviews with campaign engagers, we examine four roles of humour-based creative work in HCI: a connector to critical friends, a visceral and emotional harbour, social glue, and resistance to power. We argue for the importance of creative practices in bridging social and emotional gaps between people and concerns around technology, and in surfacing hidden perspectives for more inclusive conversations.

In the densifying data ecosystem of today's cities, data intermediaries are crucial stakeholders in facilitating data access and use. Community advocates live in these sites of social injustices and opportunities for change. Highly experienced in working with data to enact change, they offer distinctive insights on data practices and tools. This paper examines the unique perspectives that community advocates offer on data intermediaries. Based on interviews with 17 advocates working with 23 grassroots and nonprofit organizations, we propose the quality of "near" and "far" to be seriously considered in data intermediaries' works and articulate advocates' vision of connecting "near data" and "far data." To pursue this vision, we identified three pathways for data intermediaries: align data exploration with ways of storytelling, communicate context and uncertainties, and decenter artifacts for relationship building. These pathways help data intermediaries to put data feminism into practice, surface design opportunities and tensions, and raise key questions for supporting the pursuit of the Right to the City.

Advances in text-to-image generative models have made it easier for people to create art by just prompting models with text. However, creating through text leaves users with limited control over the final composition or the way the subject is represented. A potential solution is to use image prompts alongside text prompts to condition the model. To better understand how and when image prompts can improve subject representation in generations, we conduct an annotation experiment to quantify their effect on generations of abstract, concrete plural, and concrete singular subjects. We find that initial images improved subject representation across all subject types, with the most noticeable improvement in concrete singular subjects. In an analysis of different types of initial images, we find that icons and photos produced high quality generations of different aesthetics. We conclude with design guidelines for how initial images can improve subject representation in AI art.

Advances in multimodal AI have presented people with powerful ways to create images from text. Recent work has shown that text-to-image generations are able to represent a broad range of subjects and artistic styles. However, finding the right visual language for text prompts is difficult. In this paper, we address this challenge with Opal, a system that produces text-to-image generations for news illustration. Given an article, Opal guides users through a structured search for visual concepts and provides a pipeline allowing users to generate illustrations based on an article’s tone, keywords, and related artistic styles. Our evaluation shows that Opal efficiently generates diverse sets of news illustrations, visual assets, and concept ideas. Users with Opal generated two times more usable results than users without. We discuss how structured exploration can help users better understand the capabilities of human AI co-creative systems.

With the development of e-health, the number of doctors providing consultation services in online healthcare communities (OHCs) is growing. Their aim is to help patients obtain healthcare information and treatment. Since the doctors’ contributions are essential to a sustainable development of OHCs, understanding why doctors contribute to OHCs is crucial. However, the related literature that investigate motivators of doctors’ contribution behaviours in OHCs is scant. OHCs are a type of novel online community through which doctors not only obtain personal compensation but also interact with patients to build their relationship network. Hence, both personal and social motivators may affect doctors’ contributions to OHCs. Based on the theories of self-determination and Maslow’s hierarchy of needs, we established an empirical model to explore the effects of reputation, monetary rewards, doctor–patient interaction and professional status. The empirical results show that both personal and social motivators have positive effects on doctors’ contributions to OHCs, and that doctors’ professional status has a moderating effect. These findings help us understand the motivational mechanisms of doctors’ contributions to OHCs.

Instance-based learning has been proved to be effective for credit assessment in Peer-to-peer (P2P) lending. A key challenge of this application is how to measure the similarity of loans, which usually have multiple features gained from different data sources and models. In this paper, a new similarity measure is introduced to effectively integrate the information from different sources and models for credit assessment in P2P lending. Specifically, we firstly deconstructed the characteristics of P2P lending and presented four heterogeneous distance functions, which were generated by different models and information sources, to measure the loans’ similarity. Then, we proposed an integrated similarity measure that combined the above similarities by minimizing their conflicts, which could overcome the bias of the single model and single information source. Finally, we employed the portfolio selection model to develop our investment strategy. Experimental results using real datasets from Prosper demonstrated that our integrated similarity measure improves the performance of the instance-based credit assessment in P2P lending.

Point clouds provide a compact and efficient representation of 3D shapes. While deep neural networks have achieved impressive results on point cloud learning tasks, they require massive amounts of manually labeled data, which can be costly and time-consuming to collect. In this paper, we leverage 3D self-supervision for learning downstream tasks on point clouds with fewer labels. A point cloud can be rotated in infinitely many ways, which provides a rich label-free source for self-supervision. We consider the auxiliary task of predicting rotations that in turn leads to useful features for other tasks such as shape classification and 3D keypoint prediction. Using experiments on ShapeNet and ModelNet, we demonstrate that our approach outperforms the state-of-the-art. Moreover, features learned by our model are complementary to other self-supervised methods and combining them leads to further performance improvement.

The problem of pollution control has been mainly studied in the environmental economics literature where the methodology of game theory is applied for the pollution control. To the best of our knowledge this is the first time this problem is studied from the computational point of view. We introduce a new network model for pollution control and present two applications of this model. On a high level, our model comprises a graph whose nodes represent the agents, which can be thought of as the sources of pollution in the network. The edges between agents represent the effect of spread of pollution. The government who is the regulator, is responsible for the maximization of the social welfare and sets bounds on the levels of emitted pollution in both local areas as well as globally in the whole network. We first prove that the above optimization problem is NP-hard even on some special cases of graphs such as trees. We then turn our attention on the classes of trees and planar graphs which model realistic scenarios of the emitted pollution in water and air, respectively. We derive approximation algorithms for these two kinds of networks and provide deterministic truthful and truthful in expectation mechanisms. In some settings of the problem that we study, we achieve the best possible approximation results under standard complexity theoretic assumptions. Our approximation algorithm on planar graphs is obtained by a novel decomposition technique to deal with constraints on vertices. We note that no known planar decomposition techniques can be used here and our technique can be of independent interest. For trees we design a two level dynamic programming approach to obtain an FPTAS. This approach is crucial to deal with the global pollution quota constraint. It uses a special multiple choice, multi-dimensional knapsack problem where coefficients of all constraints except one are bounded by a polynomial of the input size. We furthermore derive truthful in expectation mechanisms on general networks with bounded degree.