Data visualization is a critical tool for supporting decision-making across various domains. However, its effectiveness in addressing various decision-making problems is yet to be investigated.This study systematically reviewed over 300 papers from major visualization journals and conferences and focused on 40 empirical studies. Using Franz and Kramer’s multidimensional decision-making framework, we classified and analyzed the characteristics of decision problems supported by data visualization.The findings revealed three key insights: (1) data visualization predominantly supports organizational and community-level decision-making, whereas applications at individual and family levels remain limited; (2) support has expanded from structured evaluation problems to include semi-structured and recognition-primed decision-making; and (3) challenges persist in addressing unstructured problems and small-scale contexts.These results highlight the evolving role of data visualization in supporting complex and interdependent decision scenarios. Future studies should address gaps in unstructured decision-making and investigate applications in small-scale contexts to enhance the adaptability and impact of visualization tools.

This study investigated whether emotional responses to graph color schemes differ between sexes. We conducted an online evaluation using bar charts with fourteen color schemes. The analysis revealed that some schemes induced large sex differences, whereas others induced small differences. Using entropy, we grouped the schemes by emotional concentration, revealing four categories that combined high and low concentrations with large and small sex differences. These two aspects appeared to be independent, although unified features for each group were not identified. These findings highlight the potential of emotion-aware visualization designs that consider individual differences.

Problems related to pickup and delivery are common with on-demand ride-sharing services. Passengers request rides through mobile apps, and drivers pick them up at designated locations and take them to their destinations. Although this process efficiently transports many passengers with a limited number of vehicles, it usually results in an NP-hard problem involving multiple travel requests from different locations. This paper presents an approach to simulating such services interactively by solving optimization problems to properly assign sets of requests to vehicles. Our approach successfully concatenates pickup and drop-off points on vehicle routes by enumerating their permutations. We randomly sample possible permutations of visit locations to approximately solve NP-hard routing problems within a specific time frame. Our interactive interface allows us to reassign additional on-demand travel requests to vehicles at any time, accommodating dynamic pickup and delivery issues.

Graph and network visualization supports exploration, analysis and communication of relational data arising in many domains: from biological and social networks, to transportation and powergrid systems. With the arrival of AI-based question-answering tools, issues of trustworthiness and explainability of generated answers motivate a significant new role for visualization. In the context of graphs, we see the need for visualizations that can convince a critical audience that an assertion (e. g., from an AI) about the graph under analysis is valid. The requirements for such representations that convey precisely one specific graph property are quite different from standard network visualization criteria which optimize general aesthetics and readability. In this paper, we aim to provide a comprehensive introduction to visual proofs of graph properties and a foundation for further research in the area. We present a framework that defines what it means to visually prove a graph property. In the process, we introduce the notion of a visual certificate, that is, a specialized faithful graph visualization that leverages the viewer's perception, in particular, pre-attentive processing (e. g., via pop-out effects), to verify a given assertion about the represented graph. We also discuss the relationships between visual complexity, cognitive load and complexity theory, and propose a classification based on visual proof complexity. Then, we provide further examples of visual certificates for problems in different visual proof complexity classes. Finally, we conclude the paper with a discussion of the limitations of our model and some open problems.

Graph and network visualization supports exploration, analysis and communication of relational data arising in many domains: from biological and social networks, to transportation and powergrid systems. With the arrival of AI-based question-answering tools, issues of trustworthiness and explainability of generated answers motivate a greater role for visualization. In the context of graphs, we see the need for visualizations that can convince a critical audience that an assertion about the graph under analysis is valid. The requirements for such representations that convey precisely one specific graph property are quite different from standard network visualization criteria which optimize general aesthetics and readability. In this paper, we aim to provide a comprehensive introduction to visual proofs of graph properties and a foundation for further research in the area. We present a framework that defines what it means to visually prove a graph property. In the process, we introduce the notion of a visual certificate, that is, a specialized faithful graph visualization that leverages the viewer's perception, in particular, pre-attentive processing (e. g. via pop-out effects), to verify a given assertion about the represented graph. We also discuss the relationships between visual complexity, cognitive load and complexity theory, and propose a classification based on visual proof complexity. Finally, we provide examples of visual certificates for problems in different visual proof complexity classes.

One of the key challenges faced by two-year specialized training colleges is the early detection of signs indicating student dropout and the implementation of strategies to prevent dropout through effective guidance. Given the wide variation in student traits and dropout factors, detecting these signs and determining effective guidance requires a certain level of experience. Therefore, a support system is required to enable inexperienced teachers to offer effective guidance. This study addresses the challenge of automating the recommendation of effective guidance methods for individual students, which is essential after identifying those students showing signs of potential dropout. The technical challenges involve suggesting effective guidance methods early on and maintaining stable accuracy despite year-to-year variations in student traits. To address these challenges, we developed a classification model that suggests effective guidance based on the data collected approximately six months after enrollment, which is adequate for determining grades and attendance at the end of the first year. By clustering students based on their traits and refining the teacher data accordingly, we achieved stable classification accuracy despite year-to-year fluctuations, thereby automatically providing effective guidance to approximately 80% of the students. This achievement aligns practically with the experience of the teachers.

When visually representing a rooted tree, a hierarchical drawing, a type of monotone drawing, is often used. In a layout where nodes with the same level are arranged on a horizontal straight line, the resulting shape often appears stretched horizontally. Depending on the aspect ratio of a rectangular drawing area, this can result in significant wasted space. This study presents a layout method developed to address this problem. The method maintains a monotone drawing but relaxes the criteria for arranging nodes with the same level on a horizontal straight line. Leaves with the same parent may be arranged on a folded line instead of a horizontal straight line; we call this a monotone local folding layout. Area adaptivity is introduced as a measure of spatial efficiency for a given drawing area, and the area adaptivity of the monotone local folding layout is investigated using 2500 randomly generated rooted trees as experimental data. This study also examines how area adaptivity is affected when the aspect ratio of the drawing area is changed. The experiments show that for rooted trees with many sibling leaves, the area adaptivity of the monotone local folding layout is good and is not affected by changes in the aspect ratio of the drawing area. As a result, the monotone local folding layout is an effective layout technique for rooted trees representing computer networks consisting of switches and many terminals.

Abstract Aggregation, as an operation of cartographic generalization, provides an effective means of abstracting the configuration of building features by combining them according to the scale reduction of the 2D map. Automating this design process effectively helps professional cartographers design both paper and digital maps, but finding the best aggregation result from the numerous combinations of building features has been a challenge. This paper presents a novel approach to assist cartographers in interactively designing the aggregation of building features in scale‐aware map visualization. Our contribution is to provide an appropriate set of candidates for the cartographer to choose from among a limited number of possible combinations of building features. This is achieved by collecting locally optimal solutions that emerge in the course of aggregation operations, formulated as a label cost optimization problem. Users can also explore better aggregation results by interactively adjusting the design parameters to update the set of possible combinations, along with an operator to force the combination of manually selected building features. Each cluster of aggregated building features is tightly enclosed by a concave hull, which is later adaptively simplified to abstract its boundary shapes. Experimental design examples and evaluations by expert cartographers demonstrate the feasibility of the proposed approach to interactive aggregation.

In this paper, the authors presented the challenges in modeling human recognition of deformed maps and attempted to model human recognition using deformed maps of prefectures based on the values observed in previous studies as the values that our model should imitate. First, the similarity values between the original and deformed shapes were calculated using several distance functions, and the relationship between the similarity and observed values was visually analyzed. Consequently, the dynamic time warping (DTW) distance was adopted as the distance function. The sigmoid function was used to fit the similarity values to the observed values. However, through trials, the authors concluded that a single fitting function could not adequately represent the human cognitive abilities. Therefore, a model consisting of several submodels was constructed. Each submodel fit a cluster of prefectures depending on their characteristics. To determine the appropriate number of clusters, the authors examined the errors between the model output and the observed values for different numbers of clusters. Consequently, the human deformation map recognition ability was modeled with an average error of approximately 20–26% for nine clusters. For some prefectures, the error can be reduced to approximately 10-15%.

Color is an essential factor that influences human perception, and thus, the proper selection of color sets is crucial in creating informative and appealing visual content. Furthermore, the choice of such color palettes often reflects the underlying emotional intention of creators, especially when they want to introduce specific affective styles. This paper presents a color palette recommendation system that facilitates preferred colors and affective expressions in visual content. This is accomplished by introducing non-negative tensor factorization (NTF), which extends the conventional matrix-based collaborative filtering for recommending items through ratings of multiple users. In our approach, we composed a rating tensor that constitutes the scores for colors in terms of affective factors provided by participants in the user study. With this rating tensor, we explored the meaningful relation between affective expression and color preference. Our experiments exposed that we can successfully apply a tensor-based approach to recommending convincing sets of colors in several possible cases by predicting the underlying emotional intentions in the visual content design.

Seabirds have varied movement patterns depending on their group, nesting site, or season. Therefore, seabird experts need to compare movement data under various conditions for conservation and elucidation of ecology. Visualization of movement data helps intuitive analysis. However, it is not easy for seabird experts to design suitable visual representations for comparison. The purpose of our study is to develop a tool to support the visualization process for comparison of seabird movement. There are three components to visualization techniques for comparison: Juxtaposition, Superposition, and Explicit Encodings. In our study, we aim to support comparative analysis under complex conditions using flexible Small Multiples based on juxtaposition and superposition. We designed visual representations for the comparative analysis of seabird movement data. We implemented such visual representations into a tool and confirmed its effectiveness for comparative analysis of seabird movement data.

To automate deformation that preserves mental images, an experiment was conducted to determine the conditions under which mental images are preserved using line simplification, a method of deformation. The experiment was conducted using prefectures in Japan as objective concepts and participants were asked whether they could identify a prefecture from its deformed contour and whether it matched their image of the prefecture. The results of the online web-based experiment yielded an average of 26-27 responses per prefecture. Based on the analysis of the results, the contours of prefectures are divided into those with and without local features. Furthermore, as local features, “spikes” with an area ratio above a certain level and a vertex angle below a certain level influenced the preservation of mental images.

Colors are convenient visual variables in data visualization and play an important role in data representation. However, we frequently see inappropriate color schemes in many different places. Such schemes may be due to the carelessness of the creators and their lack of knowledge of visual representations. The author of this paper has developed a tool to help understand color spaces and color differences. The tool displays cross-sections of color spaces and presents polygons representing color differences in the cross-sections. By using this tool, it is possible to understand the distortion of a color space against human visual characteristics. In this paper, the tool and the methods used in the development of the tool are introduced, and the effectiveness of the tool is discussed.

Creators can learn the art of their craft through different types of learning materials. In this study, we focus on learning materials of two-dimensional illustrations. Expressions for learning materials have various forms, many of which are posted as articles on illustration posting sites and magazines and others are posted as movies on video sharing services. They have their own advantages and disadvantages. In terms of data, illustration learning materials can be regarded as spatio-temporal data. The authors of this paper have considered applying illustration learning materials as spatio-temporal data consisting of events that occur during the creation process, and they proposed the use of learning materials in an event-based form. A viewing tool developed by the authors provides functions to display the learning materials in an event-based form using space-based keys. The effectiveness of the proposed form is also shown through an evaluation experiment.

Color is a convenient visual attribute in data visualization that performs significant roles in data representation. However, there are several situations where colors are used inappropriately to represent quantitative data. A possible reason could be that it is not easy to develop color ramps that consider color differences. This paper describes some tools designed to support the development of color ramps to represent quantitative data. The tools help develop color ramps with uniform color differences by selecting colors from a specified color path (continuous straight lines or curved lines) in a color space at equal intervals according to some color difference formula.

ChronoView is a visualization method of representing periodic features of the occurrence of events. It expresses a set of time stamps in the position on a plane. Although ChronoView offers high space efficiency, it can generate ambiguous representations. To solve this problem, glyphs have been exploited as ChronoView markers. This paper explains a user study conducted to investigate the effectiveness of the Star Glyph and Ring Glyph. The study shows that glyphs contribute to an accurate reading of temporal features. It also shows that Star Glyph and Ring Glyph have different features. It is clear that, whereas Ring Glyph dominates the time range reading, Star Glyph dominates the comparison of frequencies in the unit of time.

The consistent arrangement of map features in accordance with the map scale has recently been technically important in digital cartographic generalization. This is primarily due to the recent demand for informative mapping systems, especially for use in smartphones and tablets. However, such sophisticated generalization has usually been conducted manually by expert cartographers and thus results in a time-consuming and error-prone process. In this paper, we focus on the displacement process within cartographic generalization and formulate them as a constrained optimization problem to provide an associated algorithm implementation and its effective solution. We first identify the underlying spatial relationships among map features, such as points and lines, on each map scale as constraints and optimize the cost function that penalizes excessive displacement of the map features in terms of the map scale. Several examples are also provided to demonstrate that the proposed approach allows us to maintain consistent mapping regardless of changes to the map scale.

ChronoView is a visualization method representing periodic features of event occurrence times. It represents each event group (i.e., each set of time stamps) as a position on a plane. While ChronoView has high space efficiency, it has representational ambiguity. For example, while two similar sets must be placed at positions close to each other, two sets placed at positions close to each other need not always be similar. A three-dimensional development of ChronoView has been proposed to solve this problem. This paper presents a numerical experiment to evaluate how similarities between event groups are faithfully represented by Euclidean distance in a presentation space. The experiment shows that the three-dimensional ChronoView's faithfulness is higher than that of two-dimensional ChronoView and lower than that of three-dimensional multidimensional scaling (MDS) but comparable to that of two-dimensional MDS.

In several fields handling event data, exploring periodicity and similarity of the occurrence times of events is an important task. When we analyze event data, we often focus on temporal features of event groups, periodic features of events, and similarity of temporal features among event groups. This paper illustrates a visualization method to efficiently observe such features for multiple events. The method represents the temporal features of each event group with a glyph and represents the periodic features and their similarity with an arrangement of glyphs. To determine an arrangement of glyphs, a ChronoView graph representing periodic features of event groups is integrated with pseudo multidimensional scaling (MDS) in a 3D space. This paper also demonstrates the effectiveness of the visualization method for temporal data analysis with a use case exploring significant events from an actual dataset.

Coordinating the movements of vessels may reduce the risk of their collision. When computer simulation finds such coordination, analysts, including designers of the simulation, desire to understand the effects of such coordination. The risks can be treated as a type of spatio-temporal data. The authors have tried to visualize the difference between pre-coordination data and post-coordination data to represent such coordination considering factors such as target, time, area, direction, distance, purpose, and effect. Developed visualization techniques enable analysts to intuitively understand vessels' movements and collision risks at each time point.

ChronoView is one of the visualization methods for temporal data. It arranges points representing event groups on a circle according to a rule so that it can visualize temporal features of many event groups together. However, the rule is based on a pre-specified display period; therefore, it is not suited to exploring periodic features for unknown periods. The authors extend ChronoView to a 2.5D representation and develop a visualization tool that can analyze temporal data about various periods efficiently. The visual representation has enabled analysts to search for the display periods expressing periodicity of events by displaying several charts of ChronoView in a 3D space and emphasizing the difference in positions of the events due to the difference of the display periods. The tool has functions to assist searching by showing the spectrum and the histogram calculated from the occurrence distribution of events. To demonstrate the effectiveness of the tool, the authors describe a use case of searching for periodic features from actual event data.

Spatio-temporal dependent data, such as weather observation data, is data in which attribute values depend on both the time and space in which they are recorded. Typical visualization methods of such data that are employed in mass communication involve plotting the attribute values at each point in time on a map, and either displaying a series of such maps in time order using animation or displaying them by juxtaposing horizontally or vertically. Such methods are widely known, even by non-experts in analysis, but they have some problems. These methods force readers who want to grasp spatial changes in the attribute values to memorize the representations on the maps. The longer the time-period of data, the higher the cognitive load. In order to address such problems, we develop a novel visualization technique, named "Amatsubu," which statically represents multiple instantaneous values on a single map by overlaying them. We confirm the usefulness of this method through user studies, and also determine a weak point. The weakness is a lack of readability of information for each point in time, which can induce misreadings of spatial changes. We attempt to overcome this issue by introducing animation to Amatsubu, and transforming it into a semi-static representation technique. We confirm the effect of this improvement through another user study.

The analysis of incidents is important for preventing collisions between moving objects. However, it requires that many cases of incidents be collected as target data. The purpose of this study was to develop a method that supports the collection of many cases of anomalous (near-miss) approach from massive movement data. To achieve this, a relative superposition method is proposed. Relative superposition is a type of visualization method, in which the trajectories of pairs of moving objects are transformed, focusing on their relative positional relationships. The method is expected to facilitate the search for anomalous approach cases. This paper describes five types of relative superposition, together with a visual tool in which they are implemented. Moreover, it presents a case study in which anomalous approach cases of ships were searched by applying the tool to actual the automatic identification system (AIS) data.

A weighted free tree is an undirected connected graph with no cycles whose edges and nodes have weights (positive real numbers). The purpose of our research is to support the comparative analysis of weighted free trees. The fundamental categories of visualization techniques that support comparison are juxtaposition, superposition, and explicit encoding. Juxtaposition is often used for comparison. However, the cognitive load required to grasp a difference is large. Superposition is effective for recognizing a difference, but select information may be difficult to observe. Explicit encoding is an effective technique specialized in a specific task, but representation switching may occur frequently during analysis. Hence, there is the risk of destructing a viewer's mental map, which reduces work efficiency. We divide the working process of analysis into 14 tasks and develop four visualization techniques that can continually perform these tasks. Moreover, we conduct an experiment to investigate whether our proposed techniques are able to perform these tasks. We confirm that these tasks can be performed efficiently by applying our techniques to a case study that compares two trees.

A promising approach for preventing collision accidents between moving objects is to automatically detect a situation where the status of the objects is anomalous. To exploit machine learning techniques for detecting anomaly status, we should prepare suitable models to express the statuses of moving objects in advance. In this paper, several variations of vector models are introduced. These vector models express the relative trajectories of two moving objects approaching each other. To investigate the features of the vector models, the authors tried to observe them by using panel matrices. By means of visualizations using a panel matrix, we can easily grasp the dependence of vector models to parameters and the similarity among variations of the models.

Color plays an important role in information visualization. To use the colors as visual representations of the data effectively, designing color schemes that consider perceptual differences between colors is necessary. We developed a tool to assist in color-scheme design when using the CIE1976L*a*b* (or CIELAB) color space. The developed tool provides functions to assist in designing color schemes for nominal, ordinal, and quantitative data. The CIELAB color space has a property in which the Euclidean distances between any two colors in the space can approximate the perceptual differences between them. Thus, designing color schemes for data representation is convenient. However, because the CIELAB color space has a distorted shape, color scheme design in such space is not easy. Our tool allows the designer to see the positions of colors in the CIELAB color space visually to facilitate scheme design. Moreover, our tool provides information to the designer on the distinguishability of the selected colors, and offers functions to correct color schemes automatically.

A photogenic location is a good place to take photos. It generally has beautiful scenery, historical structures, etc. When amateur photographers plan a photo trip, they often decide beforehand on photogenic locations to visit. They collect information about the locations and consider various aspects. The aim of the authors is to assist amateur photographers in planning a photo trip. Information and procedures have been organized to help select photogenic locations from the perspective of information design. In addition, an interactive visual tool has been developed to help select photogenic locations. The tool displays information on photogenic locations extracted from a large collection of geotagged photos, provided by a photo-sharing site. The tool allows us to select locations while referring to their geographic relationships. Consequently, the process of selecting photogenic locations becomes easy.

In exploratory analysis of multivariate datasets, performing an analytical task is often necessary. Such tasks may include extracting characteristic subsets and comparing them. Therefore, we support searching and comparing features of multivariate datasets. We developed Blade Graph, which is a visualization technique for comparing distributions by emphasizing coloring according to the size of the difference. In addition, we developed a visual analysis tool with representations for comparing data distributions. In a case study of our analysis tool, we analyzed collective tendencies from a social media dataset.

Emotion-weather maps are a type of thematic map used to represent emotions. The maps represent spatial distributions of complex emotions from a large group of people. Most related work focuses on a simple sentiment and draws its distributions on a map. The proposed emotion-weather maps are a set of eight maps, each of which represents one of eight categories of emotions. Emotion data are extracted from social media. The data have some biases in some aspect of time, space, and categories of emotions. Such biases restrict the observation of relatively few emotions. To address this problem, a method to normalize data is proposed. As a use case, two sets of maps are presented. These maps represent the distribution of emotions in an actual day. An earthquake on that day caused fear and surprise in a region of Japan. The maps show the emotion distributions and changes.

Widespread use of GPS terminals has made it possible to collect geospatial movement data, and visualization is an effective method to understand such data. However, for large scale movement analysis, because the data set includes complex movements of individuals, it is difficult to understand using naive visualization methods. In order to solve this problem, we developed a visualization technique that can represent large scale movement data in an aggregate manner. This visualization technique has two representations: "amoeba representation" and "amoeba colony representation." Amoeba representation represents the distance moved from a point in any direction with map scale in a geographical space, and amoeba colony representation represents movements over a wide geographical space.

In visual analytics, data comparison is a means of analyzing data. We developed Parallel Box to support the visual analysis of multivariate data by facilitating the flexible comparison of numerous multivariate items. To compare the data distributions of multivariate data, we combine cumulative bar charts and box plots, tools that are widely used in statistics. Using shadow expression based on the visual Gestalt principles of grouping, Parallel Box enables a direct comparison between either datasets or variables. We performed a social media analysis as a case study of Parallel Box. The results of our analysis confirm that Parallel Box is useful for visual analysis.

The bag-of-features models is one of the most popular and promising approaches for extracting the underlying semantics from image databases. However, the associated image categorization based on machine learning techniques may not convince us of its validity since we cannot visually verify how the images have been classified in the high-dimensional image feature space. This paper aims at visually rearrange the images in the projected feature space by taking advantage of a set of representative features called visual words obtained using the bag-of-features model. Our main idea is to associate each image with a specific number of visual words to compose a bipartite graph, and then lay out the overall set of images using anchored map representation in which the ordering of anchor nodes is optimized through a genetic algorithm. For handling relatively large image datasets, we adaptively merge a pair of most similar images one by one to conduct the hierarchical clustering through the similarity measure based on the weighted Jaccard coefficient. Voronoi partitioning has been also incorporated into our approach so that we can visually identify the image categorization based on support vector machine. Experimental results are finally presented to demonstrate that our visualization framework can effectively elucidate the underlying relationships between images and visual words through the anchored map representation.

Owing to a limited display resolution, it may be difficult to obtain an overview of high-dimensional data in the display area used for visualization. In this paper, we aimed to obtain an overview of high-dimensional data in a limited screen area. We developed Colored Mosaic Matrix as a method to obtain a data overview. Colored Mosaic Matrix is a visualization method for high-dimensional categorical data that uses a color representation of the features. By representing quantitative data in category units, the proposed method enables the visualization of data containing a large number of records. As a result of an experimental investigation of its readability, we found our method to be useful in obtaining a data overview.

In this paper, a technique for visualizing the changing rank and value of events simultaneously with time, as well as a trend analysis tool that uses the resulting visualization, are proposed. This technique facilitates analysts' understanding of the trends of both the entire data set and each event. It also uses colors, which are estimated automatically from the events' text labels, to represent the category to which events belong. Using the tool, the authors analyzed trends of topic data extracted from Twitter as use case, and could observe the trends and overview of Twitter. In addition, using the tool it is possible to present the different ways in which people who use different media think, using media-generated data.

Treemap is a visualization method for hierarchical structures in which nodes are drawn as rectangles and arranged in a nested style. Several variations of Treemap have been developed to represent different types of data. In this paper, we propose an Edge Equalized Treemap, a representation that embeds visual data such as a bar chart in leaf rectangles. This representation is characterized by leaf rectangles of equal widths. Because their widths are equal, the scale intervals of charts in a leaf rectangle can be unified, meaning that we can compare charts simply by looking at them. We compare the Edge Equalized Treemap with existing layout methods, and demonstrate the usefulness of our approach.

This paper presents a method of visualizing data that contains temporal information, such as a human's behavior and the time at which it occurs. A feature of the data is that each event may have one or more time-stamps. By analyzing this kind of data, we are able to find some behavioral patterns and obtain knowledge applicable to many fields, such as marketing research and security. We develop ChronoView, a visualization technique to support the analysis of data with temporal information. ChronoView represents an event with a set of time-stamps as a position inside a circle, similar to the dial of an analog clock. By representing each event as a position on a two-dimensional plane, we can simultaneously visualize many events and easily compare their occurrence patterns. We implement a tool based on ChronoView, which is enriched with additional functions and overcomes the drawbacks of the original system. A use case involving tweet data from Twitter illustrates the use and practicality of ChronoView.

A bipartite graph consists of a set of nodes that can be divided into two partitions such that no edge has both endpoints in the same partition. A semi-bipartite graph is a bipartite graph with edges in one partition. Anchored map is a graph drawing technique for bipartite graphs and provides aesthetically pleasing layouts of graphs with high readability by restricting the positions of nodes in a partition. For this research, the objects of the anchored map technique were extended to semi-bipartite graphs. A hybrid layout style of anchored maps and matrix representations are proposed, and an automatic drawing technique is shown. The proposed technique arranges the nodes in one partition on a circumference like the anchored map of bipartite graphs. It also divides nodes in the other partition with edges into clusters and represents them in the matrix representations to make it easy to see connective subsets.

In recent years, music-finding services have been increasing. If we have explicit information specifying pieces of music, we can find music to our taste using such services. This paper describes a tool to support music discovery. The tool visualizes a relational structure among music genres and the music-preference data of many listeners to make the users aware of their favorite music without explicit information. A case study is described to illustrate the usefulness of the tool.

An intermediate representation between tabular forms and parallel coordinates is proposed. The new representation makes it possible to switch back and forth between tables and parallel coordinates in smooth transformations. The authors have developed a tool that can manipulate these representations for the analysis of multidimensional data. Through experimental evaluation, the advantages of this tool are illustrated. It is expected that both the representation and the tool offer a significant improvement to the early stage of data analysis.

Bipartite graphs are often used to illustrate relationships between two sets of data, such as web pages and visitors. At the same time, information is often organized hierarchically, for example, web pages are divided into directories by their contents. The hierarchical structures are useful for analyzing information. Graphs with both a bipartite structure and a hierarchical structure are called “clustered bipartite graphs.” A new clustered bipartite graphs visualization technique was developed for representing both bipartite and hierarchical structures simultaneously. In this technique, nodes in one set of the bipartite graph, which are leaves of a tree, are arranged in hierarchical multi-circular style. Then, nodes in the other set of the bipartite graph are arranged by the force-directed method. The technique enables step-by-step exploration for large-scale bipartite graphs.

Our efforts for developing and evaluating human thinking support tools are summarized. First, the organization and functions of the tools are described briefly, then three levels for evaluation are identified, and several methods in each level are explained. Results for applying the evaluation methods to the tools show the effectiveness of them.

The D-ABDUCTOR system was originally developed as a diagram-based idea organizer with facilities for visualization and manipulation of a large class of graphs. It has since been enhanced to improve its extensibility, and now function exchangeability, functional independence, and configurability without programming have become important features. The enhanced system and its features are illustrated with several examples of its application.

The rapid growth of data in large databases, such as text databases and scientific databases, requires efficient computer methods for automating analyses of the data with the goal of acquiring knowledges or making discoveries. Because the analyses of data are generally so expensive, most parts in databases remains as raw, unanalyzed primary data. Technology from machine learning (ML) will offer efficient tools for the intelligent analyses of the data using generalization ability. Generalization is an important ability specific to inductive learning that will predict unseen data with high accuracy based on learned concepts from training examples. In this article, we apply ML to text‐database analyses and knowledge acquisitions from text databases. We propose a completely new approach to the problem of text classification and extracting keywords by using ML techniques. We introduce a class of representations for classifying text data based on decision trees; (i.e., decision trees over attributes on strings) and present an algorithm for learning them inductively. Our algorithm has the following features: It does not need any natural language processing technique and it is robust for noisy data. We show that our learning algorithm can be used for automatic extraction of keywords for text retrieval and automatic text categorization. We also demonstrate some experimental results using our algorithm on the problem of classifying bibliographic data and extracting keywords in order to show the effectiveness of our approach.

An automatic method for drawing compound digraphs that contain both inclusion edges and adjacency edges are presented. In the method vertices are drawn as rectangles (areas for texts, images, etc.), inclusion edges by the geometric inclusion among the rectangles, and adjacency edges by arrows connecting them. Readability elements such as drawing conventions and rules are identified, and a heuristic algorithm to generate readable diagrams is developed. Several applications are shown to demonstrate the effectiveness of the algorithm. The utilization of curves to improve the quality of diagrams is investigated. A possible set of command primitives for progressively organizing structures within this graph formalism is discussed. The computational time for the applications shows that the algorithm achieves satisfactory performance.>