Understanding and Improving Drilled-Down Information Extraction from Online Data Visualizations for Screen-Reader Users

Data visualization has become an increasingly important means of effective data communication and has played a vital role in broadcasting the progression of COVID-19. Accessible data representations, however, have lagged behind, leaving areas of information out of reach for many blind and visually impaired (BVI) users. In this work, we sought to understand (1) the accessibility of current implementations of visualizations on the web; (2) BVI users’ preferences and current experiences when accessing data-driven media; (3) how accessible data representations on the web address these users’ access needs and help them navigate, interpret, and gain insights from the data; and (4) the practical challenges that limit BVI users’ access and use of data representations. To answer these questions, we conducted a mixed-methods study consisting of an accessibility audit of 87 data visualizations on the web to identify accessibility issues, an online survey of 127 screen reader users to understand lived experiences and preferences, and a remote contextual inquiry with 12 of the survey respondents to observe how they navigate, interpret, and gain insights from accessible data representations. Our observations during this critical period of time provide an understanding of the widespread accessibility issues encountered across online data visualizations, the impact that data accessibility inequities have on the BVI community, the ways screen reader users sought access to data-driven information and made use of online visualizations to form insights, and the pressing need to make larger strides towards improving data literacy, building confidence, and enriching methods of access. Based on our findings, we provide recommendations for researchers and practitioners to broaden data accessibility on the web.

The number of Internet users has increased tenfold since the beginning of the century up to present, especially thanks to the improvements experienced in web accessibility and the growing number of languages which online content is available in. While translation professionals are making a considerable contribution to that digital information richness, little evidence exists regarding their involvement in the achievement of a more accessible web for all. In this paper, we present the main results of the first empirical study on web accessibility conceived around a translation task. The experiment sought to particularly investigate the quality of image text alternatives produced by French translators with the help of two evaluation tools: aDesigner and Acrolinx. The assessment of their alt text proposals, carried out by seven screen reader users, suggests that using both tools helps translators to create more appropriate text alternatives than when trying to do so with only one tool or without any automated support. A more in-depth analysis of the data gathered shows that Acrolinx offers better guidance than aDesigner for translators to render images accessible.

It is not surprising that rich Internet content, such as Flash and DHTML, is some of the most pervasive content because of its visual attractiveness to the sighted majority. Such visually rich content has been causing severe accessibility problems, especially for people with visual disabilities. For Flash content, the kinds of accessibility information necessary for screen readers is not usually provided in the existing content. A typical example of such missing data is the lack of alternative text for buttons, hypertext links, widget roles, and so on. One of the major reasons is that the current accessibility framework of Flash content imposes a burden on content authors to make their content accessible. As a result, adding support for accessibility tends to be neglected, and screen reader users are left out of the richer Internet experiences.Therefore, we decided to develop an automatic accessibility transcoding system for Flash content to allow users to access a wider range of existing content, and to reduce the workload for content authors by using an automatic repair algorithm. It works as a client-side transcoding system based on the internal object model inside the Flash content. It adds and repairs accessibility information for existing Flash content, so screen readers can present more accessible information to users. Our experiment using the pilot system showed that 55% of the missing alternative texts for buttons in the tested websites could be added automatically.

Visually impaired individuals, estimated at 285 million globally, rely heavily on-screen readers for internet access. However, much of the visually available information, such as the relationship between webpage elements, does not translate well to its textual representation and must be always kept in memory, limiting contextual interactions. To address this challenge, we developed Touch Matrix Assistive Technology Navigator (TOMAT), an open-source system that works alongside screen readers to provide an interactive, audio-tactile representation of webpage structure and enable contextual interactions. Our study employed a participatory design approach, involving visually impaired users, healthcare professionals, engineers, and community organizations in co-design sessions, prototype demonstrations, and focus groups. The resulting system extracts and presents non-linear web information at multiple levels of detail, allowing users to dynamically adjust granularity and efficiently navigate and interact with web content. Participants reported that TOMAT enhanced their understanding of webpage structure and provided an intuitive complement to screen reader software. The findings suggest TOMAT has the potential to improve the internet navigation experience for visually impaired users, fostering greater independence and digital participation. To support further development and collaboration, TOMAT’s source files have been released under an open-source license.

The purpose of this usability study was to determine the most effective of three ways to display hierarchical data using the interactive data visualization software, Tableau. Often, data visualizations contain large amounts of important information that users need to be able to manipulate and interpret. Viewing hierarchical data in an interactive data visualization software like Tableau has the advantage of allowing dynamic selection of the hierarchical level of detail of results displayed. This enables improved understanding and exploration of the material. However, individuals using such software do not necessarily have knowledge of a dataset and/or the data visualization software, resulting in an inability to fully investigate data relationships. It is therefore critical that research be conducted to determine which data presentation styles promote intuitive navigation within the data visualization. A within-subjects usability study was conducted to examine the most effective of three ways to display hierarchical data within a designated area of a Tableau visualization. Three distinct visualizations of hierarchical data were randomly shown to participants. Each visualization was bordered by identical contextual information with the centrally placed hierarchical data varying. One condition showed the data relying on filters. A second condition showed the data relying on users to expand and collapse the level of detail with scrolling. A third condition showed the data as a drill-down chart that only expands the level of detail selected by the user. Metrics of user-response time, the accuracy of responses to assessment questions, the subjective rank of usability for each data visualization, and open-ended user feedback were examined. Results are discussed.

Biomedical visualizations are key to accessing biomedical knowledge and detecting new patterns in large datasets. Interactive visualizations are essential for biomedical data scientists and are omnipresent in data analysis software and data portals. Without appropriate descriptions, these visualizations are not accessible to all people with blindness and low vision, who often rely on screen reader accessibility technologies to access visual information on digital devices. Screen readers require descriptions to convey image content. However, many images lack informative descriptions due to unawareness and difficulty writing such descriptions. Describing complex and interactive visualizations, like genomics data visualizations, is even more challenging. Automatic generation of descriptions could be beneficial, yet current alt text generating models are limited to basic visualizations and cannot be used for genomics. We present AltGosling, an automated description generation tool focused on interactive data visualizations of genome-mapped data, created with the grammar-based genomics toolkit Gosling. The logic-based algorithm of AltGosling creates various descriptions including a tree-structured navigable panel. We co-designed AltGosling with a blind screen reader user (co-author). We show that AltGosling outperforms state-of-the-art large language models and common image-based neural networks for alt text generation of genomics data visualizations. As a first of its kind in genomic research, we lay the groundwork to increase accessibility in the field. The source code, examples, and interactive demo are accessible under the MIT License at https://github.com/gosling-lang/altgosling. The package is available at https://www.npmjs.com/package/altgosling.

With the rapid development of virtual reality, augmented reality, mixed reality, and artificial intelligence, exhibition space design and communication have undergone significant transformations. This study systematically reviews the research progress using the Web of Science Core Collection as the data source. Publications from 2015 to 2025 were analyzed through bibliometric and visualization tools, including R-biblioshiny, CiteSpace, and VOSviewer. The analysis examined annual publication trends, keyword co-occurrence and bursts, thematic evolution, and national and institutional collaboration networks. Results show a steady growth in studies, with peaks in 2018 and 2023, reflecting strong academic interest in immersive exhibition technologies. “Virtual reality” and “augmented reality” remain central themes, while “cultural heritage”, “exhibition design”, and “user experience” are major application domains. Emerging topics such as “artificial intelligence”, “virtual environments”, and “data visualization” highlight a shift toward intelligence-driven and experience-centered research. China and the United States lead in output and influence internationally, whereas European countries show regional collaboration. Despite these advances, limitations persist regarding intelligent applications, cross-cultural comparisons, longitudinal research, theoretical development, and sustainability concerns. By mapping knowledge structures and research frontiers, this study offers valuable insights for advancing academic inquiry and guiding innovative practices in exhibition spaces.

Collaborative visual analytics (CVA) involves sensemaking activities within teams of analysts based on coordination of work across team members, awareness of team activity, and communication of hypotheses, observations, and insights. We introduce a new type of CVA tools based on the notion of “team-first” visual analytics, where supporting the analytical process and needs of the entire team is the primary focus of the graphical user interface before that of the individual analysts. To this end, we present the design space and guidelines for team-first tools in terms of conveying analyst presence, focus, and activity within the interface. We then introduce InsightsDrive, a CVA tool for multidimensional data, that contains team-first features into the interface through group activity visualizations. This includes (1) in-situ representations that show the focus regions of all users integrated in the data visualizations themselves using color-coded selection shadows, as well as (2) ex-situ representations showing the data coverage of each analyst using multidimensional visual representations. We conducted two user studies, one with individual analysts to identify the affordances of different visual representations to inform data coverage, and the other to evaluate the performance of our team-first design with exsitu and in-situ awareness for visual analytic tasks. Our results give an understanding of the performance of our team-first features and unravel their advantages for team coordination.

Virtual reality (VR) is one of the technologies that has slowly and steadily gained popularity in the recent years. Many companies such as Facebook, HTC and Samsung have invested billions of dollars in the production of VR headsets. The analyst team of the multinational finance giant Goldman Sachs predicts that VR/AR technology will have an estimated market value of 80 billion dollars by the year 2025. Most of the applications concerned with virtual reality is related to the entertainment industry. However, some recent works have shown us that VR can be used as an effective tool for data visualization. The main focus of this paper was to evaluate a data visualization tool in WebVR for a multi-dimensional data-set based on the usability and learn-ability of the user experience. This project used the Samsung Galaxy S8 as the visualization device and VR Box along with Bluetooth controller as the equipment for viewing and interaction mechanism. The implementation was majorly carried out by collecting the data-set from an online repository called UCI Machine Learning repository and the system was implemented using WebVR and D3.js. The evaluation consisted of four sections namely: pre-test questionnaire, screen record/ think-aloud technique, post-test questionnaire and emotion card. From the evaluation, the project surmised that smart-phones have the potential to be used as a data visualization platform and it can be further improved upon with future research and additional iterations. Furthermore, this project hopes that it provides information to explore innovative inputs and creating novel UI that enables us to see the changes made to the data in real time.

Q-PIT : a populated information terrainAndy Colebourne, John A. Mariani and Thomas RoddenCooperative Systems Engineering Group, Computing Department, Lancaster University, Lancaster, LA1 4YR e-mail:jam @comp.lancs.ac.ukABSTRACTThis paper describes a system called Q-PIT. It is a prototype example of a class of systems we define as populatedinformation terrains. Within the scope of our work, we examine issues from multi-user virtual reality, data visualisation andthe use of databases to support cooperative work. We describe the use of Q-PIT as an information terrain, show how such aterrain can be generated, explored and manipulated, before considering the issues of populating such a terrain with more thanone user.Keywords : data visualisation, data exploration, computer supported cooperative work, multi-user virtual reality1. INTRODUCTIONThe growth of network availability has led to a dramatic increase in the amount of on-line information users can accessremotely. However, users often have difficulty locating the details they require due to the manner in which this information isprovided. The problem is significant and growing in importance and acuteness with the massive increase in the availability ofon-line information caused by the rapid growth of the Internet and the development of facilities to promote access to on-lineinformation sources. Users will be overwhelmed by the data made available to them unless they are provided withappropriate tools to manage the information and its navigation.Most existing interfaces allow users to issue keyword based searches for information. These systems are effective fortargeted searches when users want to find a particular item but are more problematic when users are unsure of theinformation they wish. It is all too easy to get either far too much information or none at all. To alleviate this problem anumber of hypertext based interfaces have been developed to support browsing rather than searching of information. Perhapsthe most notable example is Mosaic; a graphical interface to the World Wide Web. While promoting the browsing ofinformation and the representation of links between different pieces of information, these systems provide only limitedspatial cues. Users often feel a sense of disorientation and an inability to see the overall picture [Beheshti 92]. These systemsalso provide no indication of other users or allow users to orientate their action based on the activities of others. The WorldWide Web is a lonely place to search for information you know there are over 10 million users out there yet you never metanyone!In this paper we present the development of a system called Q-PIT which provides an interface to information stores whichsupports the spatial browsing of information and the cooperative access to shared information. Q-PIT exploits a distributedmulti-user VR platform to allow information to be presented to a community of users in terms of populated informationterrains (or PITS). PITs are 3D visualisations of abstract data, which additionally contain representations of users and theiractivities within the space. As such, they impact on a wide range of research areas and issues : multi-user virtual reality,database systems, HCI, data visualisation and computer support for cooperative work (CSCW). In essence, users are offeredthe possibility of inhabiting a world constructed from the information they are using.The approach taken in this paper is to introduce topics in a stepwise fashion, adding additional detail as further areas areraised. We begin by examining our basic premise : information terrains. Building on this introduction to PITS we consider Q-PIT in terms of the construction of the space and how users interact with this space. We then introduce the cooperativeaspects of Q-PIT prior to discussing the general utility of Q-PITS.2. INFORMATION TERRAINSThe concept of Information Terrains aims to extend access to on-line information services by allowing information to begraphically presented to users. By exploiting notions drawn from Virtual Reality (VR) the information can be laid out inthree dimensions and users can move through a terrain inspecting information for significant patterns and structures. Theunderlying philosophy of an information terrain is that they should support people in working within data as opposed tomerely with data.

This paper describes and suggests the data visualization for real time power system operation based on energy management system. Because real time power system operation performs analysis of the vast amount of on-line data, the operators need intuitive data visualization to find out useful information in the big data. Especially, in emergency situation, the data visualization is able to assist the operators in handling the crisis quickly and efficiently. Therefore, this paper aims to improve displays of output of real time power system operation by visualizing on-line big data. Through this study, we can develop improved visualization technique for real time power system operation, which has highly readable displays of output and intuitive information.

The importance of Power System Measurement in the design, Making preparations and operating electrical power transmission networks are becoming more common. Based on an energy management system, this study discusses and recommends data visualisation for real-time power system operation. Because real-time power system operation involves the processing of a massive volume of online data, operators require simple data visualisation to extract meaningful information from the large data. In particular, in an emergency situation, data visualisation can aid operators in dealing with the problem swiftly and efficiently. As a result, the goal of this research is to enhance output displays of real-time power system functioning by displaying on-line huge data. Through this research, we will be able to provide a better visualisation approach for real-time power system operation, with highly legible output displays and comprehensible information. Volatility in power transmission and distribution compels power grid management to rely more heavily on communications networks to monitor and run their systems. Malevolent actors will have a larger assault surface as a result of the increased communication. Indeed, cyber attacks on power grids have historically resulted, widespread blackouts. In this research, we look at the internet architecture underlying power grids so order to uncover the fundamental issues power grid cyber security.

Emotion detection, hence, is an important step toward making human-computer interaction a more enhanced process, where systems are made capable of identifying and responding to the emotional state of users. In fact, multimodal emotion detection systems in which both auditory and visual information are fused are emerging, and these approaches toward expressive emotional states are complementary and robust. Multimodal systems enhance the quality of interacting and, through many applications, can diagnose emotional disorders, monitor automotive safety, and improve human-robot interactions. In nature, the high-dimensional space and dynamic threats have resulted in obtaining low accuracy and high computational cost in applying the traditional models based on single-modality data. On the other hand, multimodal systems explore the synergy between audio and visual data, giving better performance and higher accuracy in inferring subtle emotional expressions. The latest improvement was done on these systems using recent advancements in transfer learning and deep learning techniques.That being said, this research Proposal devises a multimodal emotion recognition system integrating speech and face information through transfer learning for improved accuracy and robustness. Serving this purpose, the objectives of this research entail the effective comparison among different transfer-learning strategies, including the impact of pre-trained models in speech-based emotion recognition, and to introduce the role of voice activity detection in the process. Advanced neural network architectures like Spatial Transformer Networks and bidirectional LSTM in facial emotion recognition will also be tested. Early and late fusion strategies will also be used to find the best strategy for combining speech and facial data.This research will target several challenges that involve the complexity of data, balancing of the model performance-robustness balance, computational limitations, and standardization of evaluations in developing a working and robust emotion recognition system to enhance digital interaction and apply in practical areas. The goal is to create a system that oversteps the limitation of single-modality models through state-of-the-art advances in deep learning, as well as front-line improvements in transfer learning, in the manner of emotion detection performance.

ABSTRACTThis article presents the vision, implementation, and case studies of the Iowa Flood Information System (IFIS) towards the vision for next-generation decision support systems for flooding. The IFIS is an end-to-end web-based platform that incorporates various aspects of the decision-making process for flood risk management and mitigation for the State of Iowa. The IFIS provides real-time information on streams and weather conditions, incorporates advanced hydrological models for flood prediction and mapping, and several data analytics and visualization tools to support effective decision-making process. The data analytics and visualization capabilities with critical flood information in the IFIS can support flood risk management, and disaster preparedness and response efforts in Iowa. The IFIS is designed in an integrated and modular structure and serves as a prototype for a generalized decision support system for flood-related decision-making. A detailed case study is presented in the paper to demonstrate the capabilities of the system in decision-making.

Objective: The objective of the study was to quantify the nutrient intake of pregnant women in their second trimester, compare the mean nutrient intake with the estimated average requirement. We aimed to arrive at the NAR, MAR and the factors influencing the MAR. Methods: It was a longitudinal cross-sectional study done during the period of January 2023 to March 2025. Two-day dietary information was taken using the 24 hr dietary recall method. Intakes were converted to nutrients using the diet cal software. Dietary indicators such as NAR and MAR were calculated. Using the r studio student’s T test, multiple linear regression and data visualization was done.