COMPASSSession 1: Non-profit organizations and social workers
Thursday 17 August: 11:00-11:50; Auditorium
This paper presents a Smart City Hunt app, which was developed to engage diverse citizens by raising awareness, providing smart experiences and collecting data to guide strategies to transform the City of Windhoek into a smart and sustainable city. The project was done in collaboration with the UNDP accelerator lab Namibia, the City of Windhoek, and youth from informal settlements to explore alternative mechanisms of involving citizens from different backgrounds. Twenty-five participants took part in the half-day city hunt, making use of different modes of transport on their journey. The participants were guided via a location-based smart application, built with Locatify, which triggered challenges and important information such as identifying high carbon emitters and job vacancies respectively. A post-experience survey revealed that the participants recommend the use of the Smart City Hunt app for further citizen engagements, acknowledging the format of creating awareness and providing information on smart cities.
Non-profit organizations (NPOs) serve marginalized communities, such as street children. Their success highly depends on donationraising and their connections with donors, where online platforms (e.g. social media, individual websites, messaging applications, etc.) play a significant role. However, small-scale NPOs face several challenges due to their resource constraints while connecting with their donors and potential donors using existing online platforms. Therefore, we performed a mixed-method study to investigate the connectivity settings among such NPOs, donors, and potential donors. Consequently, we performed semi-structured interviews with seven NPOs working for street children and 21 current donors and conducted an online survey of 42 potential donors in a developing country (Bangladesh). The findings of our study reveal influential factors pertinent to the non-profit work contexts and gaps in connectivity among the stakeholders (small-scale NPOs, donors, and potential donors). We discover that, although having an online presence positively impacts the credibility of small-scale NPOs to the donors by introducing familiarity, yet, possessing such an online presence is challenging for the resource-constrained small-scale NPOs. We further provide several design implications for improving the connectivity settings, especially in terms of online connectivity, among the stakeholders by focusing on their essential roles and reduction of their encountered challenges.
Community-based social service organizations often face multifaceted challenges, including limited resources, inadequate staffing, funding constraints, and high demand for their services. These challenges are often exacerbated when serving vulnerable communities with complex social needs. Despite these difficulties, technology holds the potential to help bridge the service gap, enabling these organizations to respond more effectively to the diverse needs of their communities. In this paper, we conducted semi-structured interviews with 21 social service organizations in the USA that serve a marginalized community affected by poverty. Our study revealed several technological challenges that these organizations face, particularly in knowledge management and outreach efforts. Based on our findings, we offer design recommendations for empowering community-based social service organizations and the people they serve through technology. By leveraging the capabilities of technology, our study aims to promote social justice by assisting community-based organizations in better serving their communities.
Attachment is the emotional bonding between a child and a caregiver. Whether or not there is a secure attachment in early childhood has a profound life-long impact on the child. In recent years, attachment-based interventions have been developed and implemented, especially with families from low socioeconomic backgrounds. One important aspect of the program is to assess the quality of parent-child interactions through audio/video recorded at home while parent-child dyads were engaged in semi-structured interaction tasks, such as “three-bag-games”. The current practice relies on human coders to rate the videos which is a time-consuming process. Using a dataset of 220 video recordings of parent-child dyads collected at home as part of an attachment-based intervention program, we prototype a machine learning approach based on human body keypoints extracted from the posture analysis tool OpenPose and voice activity features derived from audio recordings. The results show that there are potential values in using machine learning to improve the coding efficiency of parent-child interactions. When further developed and improved, this kind of model may contribute to a new vision of AI-assisted parenting coaching support to make evidence-based interventions accessible and affordable at a large scale to children and families.
This mixed methods study investigates the experiences of healthcare workers (HCWs) along gender lines during the Covid-19 pandemic in Lahore, the second most populous city in Pakistan. In-person semi-structured interviews (𝑛=62) and researcher-administered surveys (𝑛=631) were conducted with doctors and nurses in five private and public hospitals. The findings reveal that male and female HCWs shared experiences related to increased working hours, psychological burdens, and adverse financial impacts. However, female HCWs struggled more than male HCWs, as their responsibilities at home and in the workplace increased. Additionally, more female HCWs than their male peers reported experiencing occupational stress due to transportation issues, working during pregnancy, and discriminatory attitudes of the patients toward them. Building on the results from our study, we propose several technological and policy initiatives that can be adopted by governments and organizations, especially in countries like Pakistan, where women account for most of the healthcare workforce but continue to bear a heavier burden when balancing work and family.
Session 2: Sustainability
Thursday 17 August: 11:55-12:45; Auditorium
Understanding household practices, beliefs, relationships among the members, and their preferences are often overlooked in the design of home-based interventions aiming to reduce consumption. We conducted a survey in the United Kingdom (22 responses) and a follow-up interview with 13 households to inform the design of interventions for reducing household consumption by: 1) understanding household consumption practices, and 2) identifying the concerns and challenges for household engagement with sustainability practices. Our findings highlight how the perspectives, understanding, and motives for consumption reduction actively shape household practices and their intentional and non-intentional attempts to curtail consumption. Existing non-negotiable practices led to additional household consumption and we found different strategies households use to reach a shared-decision on food and energy use or to engage in sustainable practices that vary across inter-generational family members. Based on our findings, we provide opportunities for motivating and fostering engagement with sustainable practices at home.
As part of global climate action, digital technologies are seen as a key enabler of energy efficiency savings. A popular application domain for this work is smart homes. There is a risk, however, that these efficiency gains result in rebound effects, which reduce or even overcompensate the savings. Rebound effects are well-established in economics, but it is less clear whether they also inform smart energy research in other disciplines. In this paper, we ask: to what extent have rebound effects and their underlying mechanisms been considered in computing, HCI and smart home research? To answer this, we conducted a literature mapping drawing on four scientific databases and a SIGCHI corpus. Our results reveal limited consideration of rebound effects and significant opportunities for HCI to advance this topic. We conclude with a taxonomy of actions for HCI to address rebound effects and help determine the viability of energy efficiency projects.
Rising sustainability concerns in the food industry have driven the need for innovative approaches in culinary operations. Redesigning the menus and recipes from a sustainability perspective is a promising approach to reducing restaurants’ environmental impact. Chefs, as crucial decision-makers in menu and recipe planning practices, play a vital role in promoting sustainable food services. However, the literature lacks insights into chefs’ sustainable recipe planning practices and how information and communication technologies (ICTs) could support these practices. This paper addresses this gap by conducting individual interview sessions (n=10) and recipe generation workshops (n=10) with 20 chefs in total. It reveals four dimensions of sustainable recipes (locality, seasonality, frugality, and food quality) based on semi-structured interviews. It presents a novel interactive recipe planning concept called KNOBIE, which was designed to support chefs’ sustainable recipe planning practices by using insights that gathered from the interviews. Lastly, based on an assessment of this concept through online recipe generation sessions with chefs, it provides five design implications for integrating ICTs into the sustainable menu and recipe planning practices to promote sustainable food services in restaurants.
Food waste presents a substantial challenge with significant environmental and economic ramifications, and its severity on campus environments is of particular concern. In response to this, we introduce FoodWise, a dual-component system tailored to inspire and incentivize campus communities to reduce food waste. The system consists of a data storytelling dashboard that graphically displays food waste information from university canteens, coupled with a mobile web application that encourages users to log their food waste reduction actions and rewards active participants for their efforts.
Deployed during a two-week food-saving campaign at The Hong Kong University of Science and Technology (HKUST) in March 2023, FoodWise engaged over 200 participants from the university community, resulting in the logging of over 800 daily food-saving actions. Feedback collected post-campaign underscores the system?s efficacy in elevating user consciousness about food waste and prompting behavioral shifts towards a more sustainable campus. This paper also provides insights for enhancing our system, contributing to a broader discourse on sustainable campus initiatives.
Food waste presents a substantial challenge with significant environmental and economic ramifications, and its severity on campus environments is of particular concern. In response to this, we introduce FoodWise, a dual-component system tailored to inspire and incentivize campus communities to reduce food waste. The system consists of a data storytelling dashboard that graphically displays food waste information from university canteens, coupled with a mobile web application that encourages users to log their food waste reduction actions and rewards active participants for their efforts.
Deployed during a two-week food-saving campaign at [the home university of the authors] in March 2023, FoodWise engaged over 200 participants from the university community, resulting in the logging of over 800 daily food-saving actions. Feedback collected post-campaign underscores the system’s efficacy in elevating user consciousness about food waste and prompting behavioral shifts towards a more sustainable campus. This paper also provides insights for enhancing our system, contributing to the broader discourse on sustainable campus initiatives.
Session 3: Technology applications and innovations
Thursday 17 August: 11:00-11:50; Venue 4, 5, 6
As the building ages, the wall structure may become deteriorated (e.g., wall cracks, discontinuities, and corrosion) due to the variation of the environment (i.e., temperature and humidity). Moreover, these wall cracks, discontinuities, and corrosion will affect the living comfort and coziness. As such, the wall health diagnostic becomes crucial for the safety and comfort of modern buildings. However, the existing wall health detection techniques (e.g., UWB radars, acoustic sensing, and sensor embedding techniques) are high-cost, not ubiquitous, and not robust to the variation of the environment.
In this paper, we propose VibWall, a system that can use the smartphone’s sensors (i.e., accelerometer, gyroscope, and vibrator) to detect the wall’s structural health. Specifically, the wall cracks can be detected for living safety, comfort, and coziness. Our key idea is that the smartphone’s vibration is absorbed, reflected, and propagated disparately based on the physical structure of the wall. To be specific, we employ a novel challenge-response scheme, where the challenge is a sequence of heterogeneous vibration patterns from the smartphone’s vibrator, and the responses to these vibrations are sensed by the smartphone’s gyroscope and accelerometer sensors. Then, the machine learning-based classifier (e.g., random forest classifier) will be used to discriminate between the healthy wall and the wall with cracks, discontinuities, or corrosion based on these responses. Our experimental results show good performance on the wall’s structural health detection with the wall specimen and real-world walls.
This paper analyses the suitability of a Fuzzy Logic (FL) AI-trained machine learning model to assist IoT-based smart hydroponics for deployment with South African rural subsistence farmers. To answer these questions a fuzzy logic model was developed to accurately determine and adjust outputs, based on inputs received from IoT sensors monitoring the nutrient solution and environment conditions in the hydroponics system. We conducted an initial evaluation of the FL-AI, followed by a real-world test based on an existing data set and finally conducted testing on simulated data to ensure that the AI can accurately determine crisp values to pass to a could based IoT platform access by a micro-controller which adjusts the environment and settings of the hydroponics system. This approach was followed to ensure that we respect the time and availability of our research participants. We found that the system was able to accurately determine the necessary crisp values needed by the microcontroller. We further report on the suitability of FL-AI for rural subsistence farmers as the AI is less complex, reduces energy consumption, and resource waste, and the burden of manually monitoring and adjusting hydroponics systems, and is easy to use for first-time growers. We further discuss lessons learned particularly in the South African context such as the necessity of solar power-enabled systems and local wifi networks rather than cloud-based IoT platforms.
The Universal Immunization Programme (UIP) in India has a mandate to fully vaccinate all of India’s 27 million children born annually. The vaccination doses are recorded by frontline health workers on standardized paper-based Mother and Child Protection (MCP) cards, which are manually digitized by data entry operators, resulting in poor data quality, delays, and significant time and resources. In our paper, we focus on Optical Character Recognition (OCR) based automated digitization of MCP card images captured through a smartphone application developed by us. By utilizing a standardized template for the MCP cards, which is available a-priori, we register the card images and perform OCR on the extracted region of interest (ROIs). Since the cards with curvature or torn edges had poor ROIs, we built a global-local alignment technique which first approximates the ROI using global Homography and then refines using a local Homography resulting in improved accuracy. Our pipeline gives a character level accuracy of 98.73% on our dataset, against 75.02% by Google Cloud Vision and 79.26% by Azure OCR. We also describe our field testing experience, where the digitized MCP card images were used to provide useful features on the smartphone application for health workers to conduct vaccination sessions.
In the Global South, climate change has had a significant impact, resulting in more frequent and severe weather events like droughts, floods, and storms, which are leading to crop failures, food insecurity, and job loss. Populations with limited resources are especially vulnerable. In addition, coastal infrastructure and settlements are at risk as a result of rising sea levels and coastal erosion. As a result of climate change, the marginalized communities will be further disadvantaged and existing inequalities will worsen and these effects are expected to increase in intensity in the future, emphasizing the urgency of preventing and adapting to them. Although machine learning and numerical modeling have advanced, accurate forecasting of weather remain a difficult challenge due to the complex interactions between atmospheric and oceanic variables. It is crucial to understand the relationship between weather predictors in order to improve the accuracy and reliability of forecasts.
The purpose of this research is to examine the potential of vine copulas in explaining these complex relationships in different locations. Copulas provide a way to separate the marginal distributions from the dependency structure, offering a potential solution. They provide flexible way to model the dependence structure between random variables, allowing for more accurate and realistic risk assessments and simulations. Vine copulas are based on a variety of bivariate copulas, including Gaussian, Student’s t, Clayton, Gumbel, and Frank copulas. Vine copulas, specifically, are effective in high-dimensional problems and offer a hierarchy of trees to express conditional dependence. In addition, we propose how this framework can be applied within the sub-seasonal forecasting models to enhance the prediction of different weather events or variables.
Critical infrastructure, such as roads and electricity, are core systems that enable economic development. However, these crucial systems are frequently under-monitored in developing regions, resulting in lost opportunities for growth. Recent advances in remote sensing and machine learning have enabled monitoring and measurement of infrastructure faster and more frequently than traditional methods. However, ground data is often unavailable, resulting in a disconnect between labels and remotely sensed data. Furthermore, data from industrialized regions can only sometimes be transferred to regions with sparse data due to differences in the concept of quality between regions. Additionally, inconsistency in data and the complexity of ML models can introduce bias due to learned characteristics across diverse regions, leading to inaccurate predictions and recommendations for action. In this study, we train and compare traditional neural networks and vision transformers to predict road quality from medium-resolution satellite imagery and apply them to realistic data conditions: heterogeneous temporal-spatial resolutions. The best models (vision transformers) achieve AUROC scores of 0.934 and 0.685 for binary and five-class classification tasks, respectively, exhibiting results appealing for inference in otherwise unmeasured areas. Furthermore, these experiments and results showed that proper training techniques could produce accurate models from limited, heterogeneous, and low-resolution data.
Session 4: Machine Learning and AI
Thursday 17 August: 11:55-12:45; Venue 4, 5, 6
The COVID-19 pandemic has mainstreamed human mobility data into the public domain, with research focused on understanding the impact of mobility reduction policies as well as on regional COVID-19 case prediction models. Nevertheless, current research on COVID-19 case prediction tends to focus on performance improvements, masking relevant insights about when mobility data does not help, and more importantly, why, so that it can adequately inform local decision making. In this paper, we carry out a systematic analysis to reveal the conditions under which human mobility data provides (or not) an enhancement over individual regional COVID-19 case prediction models that do not use mobility as a source of information. Our analysis – focused on US county-based COVID-19 case prediction models – shows that (1) at most, 60% of counties improve their performance after adding mobility data; (2) that the performance improvements are modest, with median correlation improvements of approximately 0.13; (3) that improvements were lower for counties with higher Black, Hispanic, and other non-White populations as well as low-income and rural populations, pointing to potential bias in the mobility data negatively impacting predictive performance; and that (4) different mobility datasets, predictive models and training approaches bring about diverse performance improvements.
A growing body of work has focused on text classification methods for detecting the increasing amount of hate speech posted online. This progress has been limited to only a select number of highly-resourced languages causing detection systems to either under-perform or not exist in limited data contexts. This is majorly caused by a lack of training data which is expensive to collect and curate in these settings. In this work, we propose a data augmentation approach that addresses the problem of lack of data for online hate speech detection in limited data contexts using synthetic data generation techniques. Given a handful of hate speech examples in a high-resource language such as English, we present three methods to synthesize new examples of hate speech data in a target language that retains the hate sentiment in the original examples but transfers the hate targets. We apply our approach to generate training data for hate speech classification tasks in Hindi and Vietnamese. Our findings show that a model trained on synthetic data performs comparably to, and in some cases outperforms, a model trained only on the samples available in the target domain. This method can be adopted to bootstrap hate speech detection models from scratch in limited data contexts. As the growth of social media within these contexts continues to outstrip response efforts, this work furthers our capacities for detection, understanding, and response to hate speech.
Disclaimer: This work contains terms that are offensive and hateful. These, however, cannot be avoided due to the nature of the work
Small and marginal farmers are unable to get a good price for their produce because of several challenges they face in market participation. Aggregation of produce via farmer cooperatives and the ability to delay sales (for non-perishable crops) to when market prices are high, has emerged as a useful strategy to improve farmer incomes. We work with a network of farmer cooperatives in India growing soyabean, and explore the potential of developing a machine learning based price forecasting and sales recommendation system that produces suggestions on the best dates when harvested soyabean crops should be sold: e.g. whether to sell right away (if prices are likely to fall in the future) or to wait (if prices are likely to rise). We present an evaluation of different methods for price forecasting and a prospect theory based method to produce sales recommendations. Experiments on historical data indicate that we can provide modest gains to farmers, and we build and field test an Android application for this purpose. Early results indicate a positive feedback. Our methods can be generalized to other agricultural commodities that can be stored for several months and help farmer cooperatives to compete effectively in agricultural markets.
Accurate and efficient record linkage methods are essential to link patients between community health worker digital health apps and an EMR system, facilitating information flow and improving coordination of care. This study presents the eTrace workflow as an illustrative example, highlighting the benefits of enhanced coordination of care for patients in antiretroviral and non-communicable disease programs in rural Neno district, Malawi. This research focuses on the following major contributions: (1) development of a machine learning-based record linkage model for electronic health information systems, (2) comparison between the machine learning-based and probabilistic approaches to record linkage and (3) a concrete evaluation of our approach on real data for the eTrace workflow. A review of the standard record linkage architecture and its application to health information exchange systems is also presented. An empirical comparison conducted of logistic regression and the Fellegi-Sunter algorithms for this use case reveals comparable results. Both classifiers demonstrate an average precision of 0.86, while logistic regression achieves a higher recall at a fixed 0.90 precision of 0.74.
Consumer products contribute to more than 75% of global greenhouse gas (GHG) emissions, primarily through indirect contributions from the supply chain. Measurement of GHG emissions associated with products is a crucial step toward quantifying the impact of GHG emission abatement actions. Life cycle assessment (LCA), the scientific discipline for measuring GHG emissions, estimates the environmental impact associated with each stage of a product from raw material extraction to its disposal. Scaling LCA to millions of products is challenging as it requires extensive manual analysis by domain experts. To avoid repetitive analysis, environmental impact factors (EIF) of common materials and products are published for use by LCA experts. However, finding appropriate EIFs for even a single product under study can require hundreds of hours of manual work, especially for complex products. We present Flamingo, an algorithm that leverages natural language machine learning (ML) models to automatically identify an appropriate EIF given a text description. A key challenge in automation is that EIF databases are incomplete. Flamingo uses industry sector classification as an intermediate layer to identify when there are no good matches in the database. On a dataset of 664 products, our method achieves an EIF matching precision of 75%.
Session 5: Climate & Environment
Thursday 17 August: 16:00-16:50; Auditorium
In order to understand the ecological impacts of climate change on bacteria communities in Swiss alpine lakes, researchers in project 2000Lakes analyzed their chemistry and biology. Within the scope of this citizen science project, an educational platform has been implemented using the data mainly collected from Wikipedia and from the research results. By presenting Swiss alpine lakes and the 2000Lakes project in an interactive way, the goal of this platform is to raise interest and promote awareness about Swiss alpine lakes to the public, and ultimately, to conserve these ecosystems by joining forces with local citizens. Volunteers were invited to use the platform and answer a survey that contains a list of questions regarding Swiss alpine lakes and a list of platform usability questions. The results were used to evaluate and improve the platform. An online crowdsourcing activity was also initiated to promote the 2000Lakes project and to complete the Swiss alpine lakes database.
For a growing economy like India, most of its energy resources are obtained through extractive processes such as mining of coal and other minerals. Mining can however have many negative social and ecological impacts if it is not well regulated. Illegal mining or inadequate reclamation of abandoned mines can amplify these impacts, emphasizing the need to develop methods that can monitor changes in the land-use patterns in and around mining sites. We develop a method using machine learning on freely available satellite data to monitor the extent of mines, and couple it with other tools to monitor deforestation, changes in built-up areas, and socio-economic development, taking place in areas in the proximity of mines. We provide evaluation results of our mining delineation classifier, a feasibility check of this suite of tools to monitor mining areas over a period of five years, and finally provide a temporal characterization study over 628 mines in India. We are in the process of enhancing these tools to eventually provide a rich set of indicators to track mining areas over time.
The interaction between elephants and their environment has profound implications for both ecology and conservation strategies. This study presents an analytical approach to decipher the intricate patterns of elephant movement in Sub-Saharan Africa, concentrating on key ecological drivers such as seasonal variations and rainfall patterns. Despite the complexities surrounding these influential factors, our analysis provides a holistic view of elephant migratory behavior in the context of the dynamic African landscape. Our comprehensive approach enables us to predict the potential impact of these ecological determinants on elephant migration, a critical step in establishing informed conservation strategies. This projection is particularly crucial given the impacts of global climate change on seasonal and rainfall patterns, which could substantially influence elephant movements in the future. The findings of our work aim to not only advance the understanding of movement ecology but also foster a sustainable coexistence of humans and elephants in Sub-Saharan Africa. By predicting potential elephant routes, our work can inform strategies to minimize human-elephant conflict, effectively manage land use, and enhance anti-poaching efforts.
This research underscores the importance of integrating movement ecology and climatic variables for effective wildlife management and conservation planning.
Air pollution adversely impacts public health. The National Capital Region (Delhi-NCR) is among the most polluted urban areas in the world. One component of air pollution is PM2.5, which accounts for around 80% of deaths due to air pollution. Solutions for lowering PM2.5 levels in Delhi have been ineffective due to their unscientific design. In this paper, we build a mixed-methods model that captures the interplay of various factors—geographical, chemical, meteorological—that contribute to the concentration of PM2.5. Using domain knowledge and KDE sampling from NASA’s GEOS-CF dataset, we identify the major sources of each of the seven constituents of PM2.5. From the 68 sources thus selected, we run the NOAA’s HYSPLIT wind dispersion model to track the movement of released particles to the sink, i.e., Delhi. Using the concentration of pollutants at the sources and by tracking their movement, we can predict the PM2.5 levels at the sink and identify polluting sources. Our model performed significantly better than the baseline fixed-effects model and captured seasonal variations in all seven constituents of PM2.5. It also uncovered the impact of polluting sources hundreds of kilometers away on the air of Delhi. Policymakers can use such a model to design data-driven policy interventions.
Forests across the world play a crucial role in the fight against the climate catastrophe as well as mass extinction that characterise the Anthropocene. However, they are also increasingly threatened by destructive human practices such as agriculture and mining, but also climate change itself. This article focuses on forests in Germany, which have been devastated in recent years by heat, drought and bark beetles. Hence, forests and associated forestry practices are in urgent need of adaptation to a different climate. Several digital applications have been developed to assist with this effort. Adaptation is complicated by the epistemological challenge of climate change, that the uncertainty of how exactly climate change will affect specific local sites, as well as future markets for forest products, poses. In this short paper we review how two applications address this uncertainty in their approach to supporting the climate adaptation of forests and draw out preliminary lessons for HCI research and design.
Session 6: Finance & Economics
Thursday 17 August: 16:55-17:45; Auditorium
Tanzanian mobile money and telecom agents (called wakala(s) in Swahili) have played a crucial role in expanding digital financial services (DFS) to rural areas. However, wakalas are losing their ability to financially sustain themselves and therefore provide compensated/uncompensated intermediation services that their communities require. This work explores the potential for the wakala network to extend intermediation services to emerging ICTs beyond the scope of commercial DFS by uncovering the social and institutional factors that currently shape wakala practices. First, we investigate how two different models of intermediation from ICTD literature can inform broader strategies for intermediation. We then complement this research with an on-the-ground quantitative survey and focus groups with community members and wakalas in Kagera, Tanzania. Our focus groups reveal that community members face challenges with new ICTs that require sustained intermediation and that wakalas encounter mounting financial instability and are thus receptive to intermediating for other ICTs. Finally, we examine three factors that influence the broadening of the wakalas‘ role of general ICT intermediaries: (1) aligning incentives and addressing the limits of pro bono actions, (3) providing appropriate training and a suitable support infrastructure, and (3) fostering trust-building and reciprocity.
This paper analyzes the current practices of Indigenous data sovereignty in environmental research and activism in the United States, as known by the settler government. The CARE principles are a widely adopted set of guidelines for Indigenous data sovereignty, yet there exists little detail on current practices of operationalization and im- plementation of the CARE principles. This research specifically iden- tified opportunities to further clarify how environmental data can
be managed in accordance with the CARE principles. Using current literature, we examine how sustainability and Human-Computer Interaction (HCI) research could better incorporate Indigenous data sovereignty and governance. Through three interviews with Indige- nous environmental practitioners, we use inductive and deductive analysis to understand current thoughts and practices. In a forestry analysis case study with the Penobscot Nation, we examine specifi- cally how the CARE principles could be implemented into a research project. The interviews and case study reveal design considerations such as emphasizing roles in responsibility and ethics to be taken into future HCI research involving Indigenous data sovereignty in environmental contexts.
A key challenge in the design of effective anti-poverty programs is determining who should be eligible for program benefits. In developing countries, one of the most common criteria is a Proxy Means Test — a rudimentary decision rule that determines eligibility based on basic information about each household (such as the number of children, whether there is indoor plumbing, etc.). At the core of each Proxy Means Test (PMT) is a machine learning algorithm, which uses the short list of household characteristics to predict whether the household should be deemed poor, and therefore eligible, or non-poor, and therefore ineligible. Using nationwide survey data from four African countries, this paper documents an important weakness in this application of machine learning: that the accuracy of the PMT prediction algorithm decreases steadily over time, by roughly 1.7 percentage points per year. We illustrate the implications of this finding for real-world anti-poverty programs, which typically update the PMT model only every 5-8 years, and then show that the aggregate effect can be decomposed into two forces: “model decay” caused by model drift, and “data decay” caused by changing household characteristics. Our final set of results show how an understanding of these forces can be used to optimize data collection policies, and how that optimization in turn can improve the effectiveness of anti-poverty programs.
Longitudinal analysis of socio-economic development at sub-national scales can reveal valuable insights about which areas tend to develop faster than others and why. Such analysis is however difficult to conduct with traditional data sources such as censuses and surveys which are not repeated frequently and may require assumptions for imputation of values at non-surveyed locations. Indicators of socio-economic development based on satellite data have emerged as a proxy to track development at fine spatial and temporal scales. We build a model using daytime and nightlights satellite data to estimate an index of socio-economic development at the village level in India. We evaluate our model for temporal robustness and use it to produce estimates at three time points over a two decade period. We then use these estimates to understand the effect on village-level development of factors such as the geographic distance of a village to hubs of economic activity and the inequality of development in the district. Our findings provide evidence of the possible impact that policy changes during this period have had on village development
This work presents an approach for combining household demographic and living standards survey questions with features derived from satellite imagery to predict the poverty rate of a region. Our approach utilizes visual features obtained from a single-step featurization method applied to freely available 10m/px Sentinel-2 surface reflectance satellite imagery. These visual features are combined with ten survey questions in a proxy means test (PMT) to estimate whether a household is below the poverty line. We show that the inclusion of visual features reduces the mean error in poverty rate estimates from 4.09% to 3.88% over a nationally representative out-of-sample test set. In addition to including satellite imagery features in proxy means tests, we propose an approach for selecting a subset of survey questions that are complementary to the visual features extracted from satellite imagery. Specifically, we design a survey variable selection approach guided by the full survey and image features and use the approach to determine the most relevant set of small survey questions to include in a PMT. We validate the choice of small survey questions in a downstream task of predicting the poverty rate using the small set of questions. This approach results in the best performance – errors in poverty rate decrease from 4.09% to 3.71%. We show that extracted visual features encode geographic and urbanization differences between regions.
Session 7: Privacy, Trust, Security, Ethics
Friday 18 August: 11:00-11:50; Auditorium
We designed the system to manage, verify and exchange identity information for Namibia’s national Digital-ID. We applied Grounded Theory methods to five focus groups to understand experiences and expectations in different contexts of legal identity verification and sharing. While local perspectives on privacy aligned with prevalent models for Digital-ID, in which people individually own and trade their personal information, they cannot be disentangled from social relations. Thus, our design responds to the ways people establish trust with organisations over time and relate consent and privacy control to organisational accountability. Our Situational Analysis considered the policy-design-adoption ‘knot’ in constructing data governance, and relations between data protection and privacy policy discourse, social structures and Namibia’s sociotechnical imaginary of ‘unity in diversity’. Our thick analysis revealed how unequal access to telecommunications contributes to systems that can produce inegalitarian and harmful data relations and prompted designing a collectivist approach to consent for information exchange that leverages government notices and civil society activism. While analysis of the policy knot improved design, our reflections also show some of the challenges in reconciling real-world design of information systems with nationwide impact in the Global Souths with best practices for Digital-ID and scholarly norms of the Global North.
Software security practices are critical in minimizing vulnerabilities and protecting unauthorized access to the code and the system. However, software security practices outside Western countries need to be better understood. This need for understanding security practices is further necessitated by the increased outsourcing of software development which can result in vulnerabilities on a global scale. This paper addresses this gap, focusing on Bangladesh, a country that represents a booming software industry in the Global South. In this study, we conducted semi-structured interviews with 15 developers to understand their security perceptions and identify the factors influencing software security practices in Bangladesh. Our findings unpack how security fits in the local software development life cycle and shed light on the challenges deterring security practices in Bangladesh. Based on our results, we provide recommendations for developing situated and sustainable strategies to support software security practices in the local context.
Despite the increase in university courses and curricula on the ethics of computing there are few studies about how CS programs should account for the diverse ways ethical dilemmas and approaches to ethics are situated in cultural, philosophical and governance systems, religions and languages. We draw on the experiences and insights of 46 university educators and practitioners in Latin America, SouthAsia, Africa, the Middle East, and Australian First Nations who participated in surveys and interviews. Our modest study seeks to prompt conversation about ethics and computing in the Global Souths and inform revisions to the ACM’s curricular guidelines for the Society, Ethics and Professionalism knowledge area in undergraduate CS programs. Participants describe frictions between static and anticipatory approaches to ethics in globalised regulations and formal Codes of ethics and professional conduct, and local practices, values and impacts of technologies in the Global Souths. Codes and regulations are instruments for international control and their gap with local realities can cause harm, despite local efforts to compensate. However, our insights also illustrate opportunities for university teaching to link more closely to priorities, actions and experiences in the Global Souths and enrich students’ education in the Global North.
The development of early career researchers (ECRs) and their induction into academia has traditionally been a process that is at best obscure, and at worst, cronyism laden. Arguably this is especially true for cross-disciplinary fields like HCI, where relatively fragmented specialisms co-exist. With COVID-19 and its negative impacts on ECRs as the backdrop, we explored the design of a five-month virtual training program for ECRs worldwide (with particular emphasis on Global South). Through an action research approach, the program was executed in collaboration with the organizers of a cross-disciplinary conference. 81 participants from 26 countries took part. The program created a collaborative learning experience for attendees and provided opportunities for networking and learning the nuances of the peer-review process. This paper details our experiences and provides reflections on design opportunities to (1) develop professional development spaces for underserved researchers, and (2) leverage ECRs’ unique capacity for contributing to inclusive conference spaces.
In the context of financial gain, hackers are motivated to exploit vulnerabilities that could result in financial or data loss. Therefore, it is crucial for financial applications to undergo thorough testing to identify and address such vulnerabilities. Regrettably, many financial institutions neglect proper testing procedures and sometimes even fail to establish a suitable security release baseline. This report presents an analysis of 18 mobile applications, each belonging to a different financial institution in Africa. The selection of these applications was carefully executed, considering institutions of varying sizes, to enable a comparative assessment of security practices across different organizational scales. The assessment was conducted by evaluating the sampled applications against the Mobile Application Security Verification Standard v2.0. This is a set of checklists and guidelines by the Open Web Application Security Project (OWASP) used as a baseline for mobile application security. Due to the extensive nature of the project, the testing scope was limited to the application itself, as experienced by the end user. This included examining the application’s interaction with the back-end server and observing its behavior on the user’s mobile device. It is important to note that this report does not provide a comprehensive analysis, as it excludes the assessment of the server-side API and testing of business logic that requires elevated privileges within the application. Furthermore, a survey was conducted to gain insights into why developers may neglect baseline security thereby introducing potential vulnerabilities in mobile applications. The findings of this survey are also included in a short summary at the end of this document.
Session 8: Social Networks and Human Factors
Friday 18 August: 11:55-12:45; Auditorium
Digital mental wellness apps are increasingly recommended for college students. Still, not all students using these apps have mood problems and do not need to be engaged in conversations involving follow-up questions. An alternate mechanism to handle such non-sensitive posts (i.e., those not indicating mood problems) could be to respond in a contextual, emotionally aware, and empathetic manner while also being terminal (not asking follow-up questions). In this paper, we evaluate the quality of training data provided by a cohort of peer college students to design AI models to respond to non-sensitive posts and minimize perceptions of being intrusive. To do so, we fine-tune the DialoGPT model using peer-provided training data, resulting in acceptable and empathetic responses with low intrusiveness (4.875%). In contrast, DialoGPT, when fine-tuned with the Empathetic Dialogue dataset, resulted in responses with high intrusiveness (69.75%), as reported by four student evaluators. We believe that mental wellness apps must be adaptive to student needs and not assume that any student posting via these apps has mood problems. The perception of intrusiveness (i.e., asking too many questions) must be considered while designing these apps. We also believe that peer students can provide a rich and reliable source of training data for college mental wellness apps.
Social media platform affordances allow users to interact with content and with each other in diverse ways. For example, on Twitter, users can like, reply, retweet, or quote another tweet. Though it’s clear that these different features allow various types of interactions, open questions remain about how these different affordances shape the conversations. We examine how two similar, but distinct conversational features on Twitter — specifically reply vs. quote — are used differently. Focusing on the polarized discourse around Robert Mueller’s congressional testimony in July 2019, we look at how these features are employed in conversations between politically aligned and opposed accounts. We use a mixed methods approach, employing grounded qualitative analysis to identify the different conversational and framing strategies salient in that discourse and then quantitatively analyzing how those techniques differed across the different features and political alignments. Our research (1) demonstrates that the quote feature is more often used to broadcast and reply is more often used to reframe the conversation; (2) identifies the different framing strategies that emerge through the use of these features when engaging with politically aligned vs. opposed accounts; (3) discusses how reply and quote features may be re-designed to reduce the adversarial tone of polarized conversations on Twitter-like platforms.
With rapid urbanization and increasing traffic congestion in major cities, alternative modes of transportation have gained significant attention. The ride-sharing app revolution also has sparked a significant transformation in the transportation industry, along with “Khep”, an unusual ride-sharing approach where individuals negotiate fares directly with drivers personally, which has become a popular means of transportation. In this research, we investigate various factors that influence the preference for such unconventional contractual rides over ride-sharing apps, such as cultural norms, trust issues, affordability, and accessibility. Moreover, we explore the role of technology literacy, marketing strategies, and regulatory frameworks in shaping the adoption landscape. After conducting a survey, we conduct a thorough analysis to determine the expected findings and uncover meaningful insights regarding the utilization, preferences, and challenges within the ride-sharing industry in Bangladesh. The findings of this study reveal that cultural factors, such as the preference for bargaining and personal connections, strongly influence the popularity of contractual rides. Additionally, concerns related to safety, data privacy, and trust act as barriers to ride-sharing app adoption. Being associated with UN SDG Goals 11 and 9, the implications of this research extend beyond Bangladesh and can provide insights for policymakers, transportation companies, and technology developers seeking to understand the factors shaping the adoption of ride-sharing apps in similar contexts.
Support plays a vital role in the teaching profession. A good support system can empower teachers to regulate their emotions and effectively manage stress while working in isolation. The COVID-19 pandemic has ushered in a hybrid form of education, necessitating the acquisition of new skills by teachers and compelling them to adapt to remote teaching. This new development further amplifies the sense of isolation prevalent amongst the teaching community. Against this backdrop, our study investigates the availability of sociotechnical support infrastructures for teachers in low-income schools while also looking into the support practices embraced by this class of teachers following the pandemic. Through 28 qualitative interviews involving teachers, management and personnel from support organizations, we demonstrate how teachers have largely taken the initiative to establish their own informal support networks in the absence of formal support infrastructures. Smartphones have significantly augmented these support practices, serving as both a valuable source of support as well as a medium for facilitating support practices. However, in comparison to other forms of support received from these sources, the availability of emotion-focused support for teachers have proven to be inadequate, creating imbalances in their support seeking practices. Our paper provides different contextual ways to reduce these imbalances and improve the occupational well-being of teachers.
The COVID-19 pandemic required handling a clear communication of risk and community engagement. A gap is noted in scholarly studies portraying strong community engagement for risk handling, particularly in resource constrained regions. This study covers community engagement and its use of technology during COVID-19 through a case study of Bangladesh. The study looks at minoritized communities who have struggled through the pandemic yet handled the difficult time through their effective problem solving, working together as a community. It is a qualitative study during the pandemic consisting of 9 communities, presenting 58 participants (N=58, Female= 33, Male=23, Transgender =2) across four divisions of Bangladesh covering urban, semi urban, and rural regions. The study uncovers the challenges and close community structures. It also shows the enhanced and increased positive role of technology during the pandemic while referring to a few communities being digitally disconnected communities that could benefit from digital connectivity in the future through increased awareness and support.
Panel 1 - The World Usability Initiative: Toward Inclusive and Usable Computing Technologies Worldwide
Elizabeth Rosenzweig, Shaima Lazem, Susan Dray
Thursday, 17 August, 16:00-17:30; Venue 4, 5, 6
Usability is the key to making future technology useful to the world, yet it is often ignored. This workshop will introduce Usability, and brainstorm ways to incorporate it into UN SDGs. One example is our effort to add World Usability Day (WUD) to the UN calendar. We will explore implications for technology design for the Global South. We will also identify concrete next steps that we all can take to advance Usability worldwide.
Panel 2 - Effective Working Environment and Factor for A Software Engineer in Companies That Are Not ICT Based
Christianah Titilope Oyewale, Akintomiwa Mayowa Abolade, Olukunle Oyewumi
Friday 18 August, 14:00-15:30; Auditorium
Working as Software Engineer in the healthcare industry for four years. There were rules guiding the work environment of healthcare professionals but none for Software engineers, who tend to overwork. Such resulted in health issues. Not ICT-based organizations need an understanding of the working factors of the Software Engineer to make them effectively produce. The panelists will discuss this and give their recommendations.
COMPASS 2023 sponsored by
