Pub Date : 2023-08-17DOI: 10.1177/14738716231189220
Kaitlyn DeValk, N. Elmqvist
Real-time situation awareness is a key challenge of cybersecurity defense. Visual analytics has been utilized for this purpose, but existing tools tend to require detailed knowledge about the network, which can be challenging in large-scale, production networks. We conducted an interview study involving 24 security professionals to gather requirements for the design, development, and evaluation of visualization to aid situation awareness in cybersecurity. Using these findings, we designed a visualization tool – called RIVERSIDE – for providing a real-time view of the dynamically changing computer network to support situation awareness. We evaluated Riverside in a user study involving 10 participants. Participants were placed in an incident response scenario that tasked them to identify malicious activity on a network. 20% of the users identified all attack component, while an additional 40% only missed one component.
{"title":"Riverside: A design study on visualization for situation awareness in cybersecurity","authors":"Kaitlyn DeValk, N. Elmqvist","doi":"10.1177/14738716231189220","DOIUrl":"https://doi.org/10.1177/14738716231189220","url":null,"abstract":"Real-time situation awareness is a key challenge of cybersecurity defense. Visual analytics has been utilized for this purpose, but existing tools tend to require detailed knowledge about the network, which can be challenging in large-scale, production networks. We conducted an interview study involving 24 security professionals to gather requirements for the design, development, and evaluation of visualization to aid situation awareness in cybersecurity. Using these findings, we designed a visualization tool – called RIVERSIDE – for providing a real-time view of the dynamically changing computer network to support situation awareness. We evaluated Riverside in a user study involving 10 participants. Participants were placed in an incident response scenario that tasked them to identify malicious activity on a network. 20% of the users identified all attack component, while an additional 40% only missed one component.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":" ","pages":""},"PeriodicalIF":2.3,"publicationDate":"2023-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48581074","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-14DOI: 10.1177/14738716231190429
Deepthi Raghunandan, Zhe Cui, Kartik Krishnan, Segen Tirfe, Shenzhi Shi, Tejaswi Darshan Shrestha, L. Battle, N. Elmqvist
Keeping abreast of current trends, technologies, and best practices in visualization and data analysis is becoming increasingly difficult, especially for fledgling data scientists. In this paper, we propose lodestar, an interactive computational notebook that allows users to quickly explore and construct new data science workflows by selecting from a list of automated analysis recommendations. We derive our recommendations from directed graphs of known analysis states, with two input sources: one manually curated from online data science tutorials, and another extracted through semi-automatic analysis of a corpus of over 6000 Jupyter notebooks. We validated Lodestar through three separate user studies: first a formative evaluation involving novices learning data science using the tool. We used the feedback from this study to improve the tool. This was followed by a summative study involving both new and returning participants from the formative evaluation to test the efficacy of our improvements. We also engaged professional data scientists in an expert review assessing the utility of the different recommendations. Overall, our results suggest that both novice and professional users find Lodestar useful for rapidly creating data science workflows.
{"title":"Lodestar: Supporting rapid prototyping of data science workflows through data-driven analysis recommendations","authors":"Deepthi Raghunandan, Zhe Cui, Kartik Krishnan, Segen Tirfe, Shenzhi Shi, Tejaswi Darshan Shrestha, L. Battle, N. Elmqvist","doi":"10.1177/14738716231190429","DOIUrl":"https://doi.org/10.1177/14738716231190429","url":null,"abstract":"Keeping abreast of current trends, technologies, and best practices in visualization and data analysis is becoming increasingly difficult, especially for fledgling data scientists. In this paper, we propose lodestar, an interactive computational notebook that allows users to quickly explore and construct new data science workflows by selecting from a list of automated analysis recommendations. We derive our recommendations from directed graphs of known analysis states, with two input sources: one manually curated from online data science tutorials, and another extracted through semi-automatic analysis of a corpus of over 6000 Jupyter notebooks. We validated Lodestar through three separate user studies: first a formative evaluation involving novices learning data science using the tool. We used the feedback from this study to improve the tool. This was followed by a summative study involving both new and returning participants from the formative evaluation to test the efficacy of our improvements. We also engaged professional data scientists in an expert review assessing the utility of the different recommendations. Overall, our results suggest that both novice and professional users find Lodestar useful for rapidly creating data science workflows.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"1 1","pages":""},"PeriodicalIF":2.3,"publicationDate":"2023-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42209321","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-11DOI: 10.1177/14738716231189218
João Moreira, Daniel Mendes, Daniel Gonçalves
Incidental visualizations are meant to be perceived at-a-glance, on-the-go, and during short exposure times, but are not seen on demand. Instead, they appear in people’s fields of view during an ongoing primary task. They differ from glanceable visualizations because the information is not received on demand, and they differ from ambient visualizations because the information is not continuously embedded in the environment. However, current graphical perception guidelines do not consider situations where information is presented at specific moments during brief exposure times without being the user’s primary focus. Therefore, we conducted a crowdsourced user study with 99 participants to understand how accurate people’s incidental graphical perception is. Each participant was tested on one of the three conditions: position of dots, length of lines, and angle of lines. We varied the number of elements for each combination and the display time. During the study, participants were asked to perform reproduction tasks, where they had to recreate a previously shown stimulus in each. Our results indicate that incidental graphical perception can be accurate when using position, length, and angles. Furthermore, we argue that incidental visualizations should be designed for low exposure times (between 300 and 1000 ms).
{"title":"Incidental graphical perception: How marks and display time influence accuracy","authors":"João Moreira, Daniel Mendes, Daniel Gonçalves","doi":"10.1177/14738716231189218","DOIUrl":"https://doi.org/10.1177/14738716231189218","url":null,"abstract":"Incidental visualizations are meant to be perceived at-a-glance, on-the-go, and during short exposure times, but are not seen on demand. Instead, they appear in people’s fields of view during an ongoing primary task. They differ from glanceable visualizations because the information is not received on demand, and they differ from ambient visualizations because the information is not continuously embedded in the environment. However, current graphical perception guidelines do not consider situations where information is presented at specific moments during brief exposure times without being the user’s primary focus. Therefore, we conducted a crowdsourced user study with 99 participants to understand how accurate people’s incidental graphical perception is. Each participant was tested on one of the three conditions: position of dots, length of lines, and angle of lines. We varied the number of elements for each combination and the display time. During the study, participants were asked to perform reproduction tasks, where they had to recreate a previously shown stimulus in each. Our results indicate that incidental graphical perception can be accurate when using position, length, and angles. Furthermore, we argue that incidental visualizations should be designed for low exposure times (between 300 and 1000 ms).","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":" ","pages":""},"PeriodicalIF":2.3,"publicationDate":"2023-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46344190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-20DOI: 10.1177/14738716231180892
João Moreira, Daniel Mendes, Daniel Gonçalves
Incidental visualizations are meant to be seen at-a-glance, on-the-go, and during short exposure times. They will always appear side-by-side with an ongoing primary task while providing ancillary information relevant to those tasks. They differ from glanceable visualizations because looking at them is never their major focus, and they differ from ambient visualizations because they are not embedded in the environment, but appear when needed. However, unlike glanceable and ambient visualizations that have been studied in the past, incidental visualizations have yet to be explored in-depth. In particular, it is still not clear what is their impact on the users’ performance of primary tasks. Therefore, we conducted an empirical online between-subjects user study where participants had to play a maze game as their primary task. Their goal was to complete several mazes as quickly as possible to maximize their score. This game was chosen to be a cognitively demanding task, bound to be significantly affected if incidental visualizations have a meaningful impact. At the same time, they had to answer a question that appeared while playing, regarding the path followed so far. Then, for half the participants, an incidental visualization was shown for a short period while playing, containing information useful for answering the question. We analyzed various metrics to understand how the maze performance was impacted by the incidental visualization. Additionally, we aimed to understand if working memory would influence how the maze was played and how visualizations were perceived. We concluded that incidental visualizations of the type used in this study do not disrupt people while they played the maze as their primary task. Furthermore, our results strongly suggested that the information conveyed by the visualization improved their performance in answering the question. Finally, working memory had no impact on the participants’ results.
{"title":"Impact of incidental visualizations on primary tasks","authors":"João Moreira, Daniel Mendes, Daniel Gonçalves","doi":"10.1177/14738716231180892","DOIUrl":"https://doi.org/10.1177/14738716231180892","url":null,"abstract":"Incidental visualizations are meant to be seen at-a-glance, on-the-go, and during short exposure times. They will always appear side-by-side with an ongoing primary task while providing ancillary information relevant to those tasks. They differ from glanceable visualizations because looking at them is never their major focus, and they differ from ambient visualizations because they are not embedded in the environment, but appear when needed. However, unlike glanceable and ambient visualizations that have been studied in the past, incidental visualizations have yet to be explored in-depth. In particular, it is still not clear what is their impact on the users’ performance of primary tasks. Therefore, we conducted an empirical online between-subjects user study where participants had to play a maze game as their primary task. Their goal was to complete several mazes as quickly as possible to maximize their score. This game was chosen to be a cognitively demanding task, bound to be significantly affected if incidental visualizations have a meaningful impact. At the same time, they had to answer a question that appeared while playing, regarding the path followed so far. Then, for half the participants, an incidental visualization was shown for a short period while playing, containing information useful for answering the question. We analyzed various metrics to understand how the maze performance was impacted by the incidental visualization. Additionally, we aimed to understand if working memory would influence how the maze was played and how visualizations were perceived. We concluded that incidental visualizations of the type used in this study do not disrupt people while they played the maze as their primary task. Furthermore, our results strongly suggested that the information conveyed by the visualization improved their performance in answering the question. Finally, working memory had no impact on the participants’ results.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"22 1","pages":"307 - 322"},"PeriodicalIF":2.3,"publicationDate":"2023-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46974064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-15DOI: 10.1177/14738716231181545
H. Alzahrani, S. Fernstad
Network biology has become crucial to understanding the complex structural characteristics of biological systems. Consequently, advanced visualization approaches are needed to support the investigation of such structures, and several network visualization tools have subsequently been developed to help researchers analyze intricate biological networks. While these tools support a range of analytical and interactive features, it is sometimes unclear to a data analyst or visualization designer which features are of most relevance to biologists. Thus, this study investigates and identifies essential factors for the visualization of complex biological networks using a mixed methodology approach. Based on the findings, essential factors were categorized as either generic and heuristic, where the former concern different analytical and interactive functionalities, such as an efficient layout, advanced search capabilities, plugin availability, graph analysis and user-friendliness, while the latter concern usability, such as information coding, flexibility, orientation and help.1 Furthermore, the findings indicate that 12 of the 15 generic factors identified were moderately important, while all 10 heuristic factors identified herein were moderately important.
{"title":"An investigation into various visualization tools for complex biological networks","authors":"H. Alzahrani, S. Fernstad","doi":"10.1177/14738716231181545","DOIUrl":"https://doi.org/10.1177/14738716231181545","url":null,"abstract":"Network biology has become crucial to understanding the complex structural characteristics of biological systems. Consequently, advanced visualization approaches are needed to support the investigation of such structures, and several network visualization tools have subsequently been developed to help researchers analyze intricate biological networks. While these tools support a range of analytical and interactive features, it is sometimes unclear to a data analyst or visualization designer which features are of most relevance to biologists. Thus, this study investigates and identifies essential factors for the visualization of complex biological networks using a mixed methodology approach. Based on the findings, essential factors were categorized as either generic and heuristic, where the former concern different analytical and interactive functionalities, such as an efficient layout, advanced search capabilities, plugin availability, graph analysis and user-friendliness, while the latter concern usability, such as information coding, flexibility, orientation and help.1 Furthermore, the findings indicate that 12 of the 15 generic factors identified were moderately important, while all 10 heuristic factors identified herein were moderately important.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"22 1","pages":"323 - 339"},"PeriodicalIF":2.3,"publicationDate":"2023-06-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44908611","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Financial institutions use credit Scoring models to predict the default of their customers and assist in decision-making about the granting of credit. As a large volume of credit transactions is generated daily alongside a potential increase in this information with the advent of Open Finance, it is challenging to monitor this information quickly so we can act in case these models lose performance. Considering this context, our research aims to provide a Visual Analytics approach to assist in monitoring credit models. For this, initially, we carried out a systematic review of the literature on the subject and conducted semi-structured interviews with 13 domain experts. Considering the needs raised with this study, we created a prototype called Visual Analytics for monitoring Credit Scoring models (VACS). The main contributions of this work are twofold: The requirements gathered through interviews with specialists, which allowed the analysis of how the models are monitored within financial institutions, something that is not disclosed and that can help in the standardization of the monitoring process; and VACS, which was evaluated by four domain experts who considered it a very complete and easy-to-use tool.
{"title":"Visual analytics for monitoring credit scoring models","authors":"Daiane Rodrigues Baldo, Murilo Santos Regio, Isabel Harb Manssour","doi":"10.1177/14738716231180803","DOIUrl":"https://doi.org/10.1177/14738716231180803","url":null,"abstract":"Financial institutions use credit Scoring models to predict the default of their customers and assist in decision-making about the granting of credit. As a large volume of credit transactions is generated daily alongside a potential increase in this information with the advent of Open Finance, it is challenging to monitor this information quickly so we can act in case these models lose performance. Considering this context, our research aims to provide a Visual Analytics approach to assist in monitoring credit models. For this, initially, we carried out a systematic review of the literature on the subject and conducted semi-structured interviews with 13 domain experts. Considering the needs raised with this study, we created a prototype called Visual Analytics for monitoring Credit Scoring models (VACS). The main contributions of this work are twofold: The requirements gathered through interviews with specialists, which allowed the analysis of how the models are monitored within financial institutions, something that is not disclosed and that can help in the standardization of the monitoring process; and VACS, which was evaluated by four domain experts who considered it a very complete and easy-to-use tool.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135672678","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-30DOI: 10.1177/14738716231173730
Antoine Béland, Florent Daudens, Thomas Hurtut
Understanding and consuming public budget data is a key issue, helping citizens in gaining insight into their democratic and political systems. The goal of this work is to present Waffster, a user-friendly representation supporting the understanding of such data. The proposed representation enables the browsing, searching, comparing, and presenting of the hierarchically arranged components and quantities in budgets. In this paper, we first conduct a thorough survey of online public budget visualizations. Then, in collaboration with Le Devoir, a Canadian daily newspaper, we propose a novel unit-based hierarchical design based on waffle charts. We evaluate this design using a controlled user study to compare it to a tree-map based layout, and a case study conducted with Le Devoir during the provincial election campaign in Québec of 2018.
{"title":"Waffster: Hierarchical waffle charts for budget visualization","authors":"Antoine Béland, Florent Daudens, Thomas Hurtut","doi":"10.1177/14738716231173730","DOIUrl":"https://doi.org/10.1177/14738716231173730","url":null,"abstract":"Understanding and consuming public budget data is a key issue, helping citizens in gaining insight into their democratic and political systems. The goal of this work is to present Waffster, a user-friendly representation supporting the understanding of such data. The proposed representation enables the browsing, searching, comparing, and presenting of the hierarchically arranged components and quantities in budgets. In this paper, we first conduct a thorough survey of online public budget visualizations. Then, in collaboration with Le Devoir, a Canadian daily newspaper, we propose a novel unit-based hierarchical design based on waffle charts. We evaluate this design using a controlled user study to compare it to a tree-map based layout, and a case study conducted with Le Devoir during the provincial election campaign in Québec of 2018.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135643711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-23DOI: 10.1177/14738716231173731
Eric Newburger, Niklas Elmqvist
We present results from a preregistered and crowdsourced user study where we asked members of the general population to determine whether two samples represented using different forms of data visualizations are drawn from the same or different populations. Such a task reduces to assessing whether the overlap between the two visualized samples is large enough to suggest similar or different origins. When using idealized normal curves fitted on the samples, it is essentially a graphical formulation of the classic Student’s t-test. However, we speculate that using more sophisticated visual representations, such as bar histograms, Wilkinson dot plots, strip plots, or Tukey boxplots will both allow people to be more accurate at this task as well as better understand its meaning. In other words, the purpose of our study is to explore which visualization best scaffolds novices in making graphical inferences about data. However, our results indicate that the more abstracted idealized bell curve representation of the task yields more accuracy.
{"title":"Comparing overlapping data distributions using visualization","authors":"Eric Newburger, Niklas Elmqvist","doi":"10.1177/14738716231173731","DOIUrl":"https://doi.org/10.1177/14738716231173731","url":null,"abstract":"We present results from a preregistered and crowdsourced user study where we asked members of the general population to determine whether two samples represented using different forms of data visualizations are drawn from the same or different populations. Such a task reduces to assessing whether the overlap between the two visualized samples is large enough to suggest similar or different origins. When using idealized normal curves fitted on the samples, it is essentially a graphical formulation of the classic Student’s t-test. However, we speculate that using more sophisticated visual representations, such as bar histograms, Wilkinson dot plots, strip plots, or Tukey boxplots will both allow people to be more accurate at this task as well as better understand its meaning. In other words, the purpose of our study is to explore which visualization best scaffolds novices in making graphical inferences about data. However, our results indicate that the more abstracted idealized bell curve representation of the task yields more accuracy.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135184672","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1177/14738716231168671
Archit Rathore, Yichu Zhou, Vivek Srikumar, Bei Wang
Transformer-based language models such as BERT and its variants have found widespread use in natural language processing (NLP). A common way of using these models is to fine-tune them to improve their performance on a specific task. However, it is currently unclear how the fine-tuning process affects the underlying structure of the word embeddings from these models. We present TopoBERT, a visual analytics system for interactively exploring the fine-tuning process of various transformer-based models – across multiple fine-tuning batch updates, subsequent layers of the model, and different NLP tasks – from a topological perspective. The system uses the mapper algorithm from topological data analysis (TDA) to generate a graph that approximates the shape of a model’s embedding space for an input dataset. TopoBERT enables its users (e.g. experts in NLP and linguistics) to (1) interactively explore the fine-tuning process across different model-task pairs, (2) visualize the shape of embedding spaces at multiple scales and layers, and (3) connect linguistic and contextual information about the input dataset with the topology of the embedding space. Using TopoBERT, we provide various use cases to exemplify its applications in exploring fine-tuned word embeddings. We further demonstrate the utility of TopoBERT, which enables users to generate insights about the fine-tuning process and provides support for empirical validation of these insights.
{"title":"TopoBERT: Exploring the topology of fine-tuned word representations","authors":"Archit Rathore, Yichu Zhou, Vivek Srikumar, Bei Wang","doi":"10.1177/14738716231168671","DOIUrl":"https://doi.org/10.1177/14738716231168671","url":null,"abstract":"Transformer-based language models such as BERT and its variants have found widespread use in natural language processing (NLP). A common way of using these models is to fine-tune them to improve their performance on a specific task. However, it is currently unclear how the fine-tuning process affects the underlying structure of the word embeddings from these models. We present TopoBERT, a visual analytics system for interactively exploring the fine-tuning process of various transformer-based models – across multiple fine-tuning batch updates, subsequent layers of the model, and different NLP tasks – from a topological perspective. The system uses the mapper algorithm from topological data analysis (TDA) to generate a graph that approximates the shape of a model’s embedding space for an input dataset. TopoBERT enables its users (e.g. experts in NLP and linguistics) to (1) interactively explore the fine-tuning process across different model-task pairs, (2) visualize the shape of embedding spaces at multiple scales and layers, and (3) connect linguistic and contextual information about the input dataset with the topology of the embedding space. Using TopoBERT, we provide various use cases to exemplify its applications in exploring fine-tuned word embeddings. We further demonstrate the utility of TopoBERT, which enables users to generate insights about the fine-tuning process and provides support for empirical validation of these insights.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"22 1","pages":"186 - 208"},"PeriodicalIF":2.3,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42372953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-20DOI: 10.1177/14738716231167181
Yan Chao Wang, Yi Xing, J. Zhang
The ordinary Voronoi treemap generated based on the Euclidean distance function has the flexibility to slightly adjust the layout when visualizing time-varying hierarchical data for better visual quality, preserving neighborhood relationships, and preserving a stable layout. However, its layout formed by segments with arbitrary orientations has poor shape stability between successive layouts at different time indexes, which is not conducive for the users to understand the plot and track the same node. In this paper, we propose novel Voronoi treemaps in Manhattan distance and Chebyshev distance respectively, such that the segments in the new layouts only have four orientations (horizontal, vertical, and ±45° to the x -axis). The new layouts can not only heritage the abilities of ordinary Voronoi treemap, but preserve good shape stability. To achieve this, we first focus on the weighted bisector between two sites in Manhattan distance and design a bisector generation method for different weight values of two sites, as the bisector is the foundation to form a diagram. Then a divide-and-conquer method is utilized to form the bisectors into a Voronoi diagram, and a Voronoi treemap layout can be finally obtained by using Lloyd’s method to iteratively adjust the diagram. Moreover, we prove that the treemap algorithm in Manhattan distance can be adjusted to also generate the Voronoi treemap in Chebyshev distance via linear transformation, avoiding designing additional algorithm. The computational properties of the proposed methods are first evaluated to check whether the layouts can be generated fast and accurately. Then the perceptual properties are evaluated quantitatively and qualitatively based on quality metrics and user studies, respectively. The results demonstrate that the proposed Voronoi treemaps preserve similar layout stability, but better visual quality and shape stability than the ordinary Voronoi treemap. Our algorithms are simple and resolution-independent, but also provide alternatives to the Voronoi treemaps.
{"title":"Voronoi treemap in Manhattan distance and Chebyshev distance","authors":"Yan Chao Wang, Yi Xing, J. Zhang","doi":"10.1177/14738716231167181","DOIUrl":"https://doi.org/10.1177/14738716231167181","url":null,"abstract":"The ordinary Voronoi treemap generated based on the Euclidean distance function has the flexibility to slightly adjust the layout when visualizing time-varying hierarchical data for better visual quality, preserving neighborhood relationships, and preserving a stable layout. However, its layout formed by segments with arbitrary orientations has poor shape stability between successive layouts at different time indexes, which is not conducive for the users to understand the plot and track the same node. In this paper, we propose novel Voronoi treemaps in Manhattan distance and Chebyshev distance respectively, such that the segments in the new layouts only have four orientations (horizontal, vertical, and ±45° to the x -axis). The new layouts can not only heritage the abilities of ordinary Voronoi treemap, but preserve good shape stability. To achieve this, we first focus on the weighted bisector between two sites in Manhattan distance and design a bisector generation method for different weight values of two sites, as the bisector is the foundation to form a diagram. Then a divide-and-conquer method is utilized to form the bisectors into a Voronoi diagram, and a Voronoi treemap layout can be finally obtained by using Lloyd’s method to iteratively adjust the diagram. Moreover, we prove that the treemap algorithm in Manhattan distance can be adjusted to also generate the Voronoi treemap in Chebyshev distance via linear transformation, avoiding designing additional algorithm. The computational properties of the proposed methods are first evaluated to check whether the layouts can be generated fast and accurately. Then the perceptual properties are evaluated quantitatively and qualitatively based on quality metrics and user studies, respectively. The results demonstrate that the proposed Voronoi treemaps preserve similar layout stability, but better visual quality and shape stability than the ordinary Voronoi treemap. Our algorithms are simple and resolution-independent, but also provide alternatives to the Voronoi treemaps.","PeriodicalId":50360,"journal":{"name":"Information Visualization","volume":"22 1","pages":"246 - 264"},"PeriodicalIF":2.3,"publicationDate":"2023-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45232514","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：