ChapterPDF Available

Cognitive Approach in Classification Visualization: end-users study

  • College of Social and Media Culture, Toruń, Poland

Abstract and Figures

Visualization of scientifi c information extends the possibility to explore how the science is organized and change over time. Particularly classifi ed data include a great potential of discovering the structure and dynamics of specifi ed domain. The authors applied tested and previously presented conception of ACM CCS (Computing Classification System) classification mapping into a sphere surface. This visualization layout could be mainly used as a multiperspective analytical tool of classification structure. Classification sphere also might be considered as an ergonomic interface for exploring scientifi c resources as well as information retrieval. Classifi ed documents form patterns according to their semantic similarity. Obtained graphical representations deliver quantitative material for analysis of classifi cation development and dynamics. The authors try to fi nd reliable tools to evaluate this spherical visualization design. They constructed an appropriate interface and surveyed the distinct groups of users, who were asked about key aspects of visualization layout. Results of our study allow evaluating classifi cation visualization thereby improving proposed methodology as well as discovering new semantic features and laws in visual layout.
Content may be subject to copyright.
Veslava Osinska
Nicolaus Copernicus University, Torun, Poland
Joanna Dreszer-Drogorob
Nicolaus Copernicus University, Torun, Poland
Grzegorz Osinski
College of Social and Medial Culture, Torun, Poland
Michal Gawarkiewicz
Nicolaus Copernicus University, Torun, Poland
Cognitive Approach in Classification Visualization:
end-users study
Abstract: Visualization of scientific information extends the possibility to explore how the science is
organized and does change over the time. Particularly classified data include a great potential of
discovering the structure and dynamics of specified domain. The authors applied tested and
previously presented conception of ACM CCS (Computing Classification System) classification
mapping into a sphere surface. Classified documents form pattern according to their semantic
similarity. Two main goals of obtained visualizations were determined. It could be mainly used as
multiperspective analytical tool of original classification and its structure. Classification sphere also
might be considered as an ergonomic interface for exploring scientific resources as well as
information retrieval. Obtained graphical representations deliver quantitative material for analysis of
classification development and dynamics. The authors try to find reliable tools to evaluate it. They
constructed an appropriate interface and surveyed the distinct groups of users, who were asked
about key aspects of visualization layout and their changes. Results of our study allow to evaluate
visualization of classification thereby to improve proposed methodology as well as to discover a
new semantic features and laws in visual layout.
Keywords: visualization interface; Infoviz; mapping; classification scheme; ACM Computing
Classification System.
1. Introduction
Information Visualization is one of the stages in data analysis process and
delivers research material in a graphical form. Data correlations and hidden
structure can be discovered using visualization layouts with more or less
complex distribution of nodes. High perceptible attributes like colour, size and
location describe the main properties of data. Is interpretation of such graphical
pattern always coherent and correct? In any case, the author(s) take main part
in interpretation, and therefore, the final conclusion involves subjective aspects.
The question Do you see the same as me?” relates to the problem of
evaluation in visual analysis. Computer scientists narrow down the evaluation
of visualization to comparison of layouts created by means of different metrics
and/or mapping algorithms (Boyack, 2005; Börner, 2010). However, this
approach does not resolve the problem of subjective interpretation because,
again, the end-analyst is a person who knows the main research problem and
may conduct his or her own scheme of assessment (Szelag et al, 2010).
Meteorological maps are claimed to be designed for a large group of recipients.
Usually cognitive mechanisms are implemented there, thus represented data
are easily understandable for all users. Examples of science maps (
Purpose and Goals 2007
) are helpful in stydying pattern recognition by human.
Authors observation of the participants of Places@Spaces exhibition leads to
the conclusion, that computer scientists take more interest in scientific data
landscapes than Humanities and Social Sciences specialists. Probably it
caused by habituation of visual communication among information technology
practicians (Carretié et all, 2003).
Birger Hjørland underlined the role of user-based study to Knowledge
Organization (2013). Cognitive approach has a long tradition (since 1970s) in
library and information science study (Birger 2013). In the past two decades
user-centered tendency became influential in broader society branches, for
example: technologies, businesses organizations, market and education.
According to Birger Hjørland Knowledge Organization, research must include
subjective aspects, and therefore is based on “collective views in discourse
communities” (Birger 2013).
This paper is a continuation of the study on visualization interface of classified
articles using a new topology. Obtained visualizations of computer science
classification and related classified articles are described from different
epistemological perspectives in a series of papers (Osinska, Bala 2010;
Osinska 2010). The authors discern two main practical applications of this
novel approach. The first one is a visual analysis of classification which
consists of scheme evolution and dynamics study as well as their consequent
improvement. The second potential concerns documents navigation space and
semantic retrieval. Because of specialization of the interface, the authors focus
on the first issue i.e. visual analysis of classification and its reliable evaluation.
In the case of classification visualization where a high level of abstraction is
involved, interpretation of graphical distribution of classes is extremely difficult.
The authors have decided to include different groups of users in the
interpretation process.
2. Model of Classification Visualization
The main problem in the graphical presentation is a low dimension topology
accessible in a standard computer monitor. Multidimensional data structure is
displayed on a plane, however the human perception systems are naturally
adapted for spherical vision (Ware, 2004). The target visualization space to be
chosen is the sphere surface because of its good ergonomic properties.
Theoretical basis of dimension reduction of large-scale data was proposed by
Thurston (1997). The analysis method applied by the author tries to exploit the
human perception and cognition mechanisms through 3D computer graphics
and interactive interface. This combination leads to natural and efficient
human-computer interaction. Such property of sphere surface as
homeomorphism preserves cohesion in terms of topology. Similar topological
space is used in visualization like Large Map of Science (Klavans & Boyack,
206) and Circos application dedicated to comparative genomics (Krzywinski,
2009). Currently, it is a standard visualization model for considerable data sets.
In the presented study, a well-known cognitive problem of Cartesians
coordinates was eliminated by using nonlinear metrics (Osinska & Bala, 2010).
Datasets constituted documents classified by Computing Classification System
and derived from Association for Computing Machinery Digital Library. The
metadata such as theme categories (classes), title, abstract, subject
descriptors and keywords were used. An original conception was based on
similarity of co-classes and the assumption: similarity degree is proportional to
the number of common publications (Osinska & Bala, 2010). The final number
of all classes and subclasses in collection determined the dimension of co-
classes similarity matrix. To place the nodes of classes and documents on
sphere surface, the multidimensional scaling technique was applied.
Documents nodes formed a pattern according to their thematic similarity. The
nodes were indicated by a proper colour of main class they belonged to
(Figure 1). The authors investigated similar and distinct research fields as well
as clusters organization by means of obtained graphical patterns. They also
analyzed the dynamics of classification due to data series for different
publishing periods with a 10-year step. The results show that visualization of
classified documents reveals both organization of digital library content and
allows to identify hierarchical thematic categories.
Figure 1: Visualization sphere – screenshot of interface for on-line classification
exploration (Application is accessible online:
3. Methodology
3.1. Application
Figure 1 demonstrates the screenshot of the application interface. The user is
able to rotate the sphere, select different combinations of main classes and
thus analyze the graphical distribution of documents nodes. The application is
accessible online in two language versions: Polish and English.
For the interface design, popular Web technologies such as Hyper Text Markup
Language with Cascading Style Sheets and JavaScript are used. This ensures
compatibility with all web browsers, but it is recommended to access the
interface using up-to-date versions of modern web browsers like Mozilla Firefox
or Google Chrome. Rendering of 3D task is provided by “Canvas K3D library”.
3.2. Survey research
Seventeen volunteers were divided into three groups (two students groups
participants ranging in age from 19 to 26 years, and one experts group),
distinguished according to subjects computer usage experience. The first
group included seven undergraduate university students who had been
recruited from different humanistic or social sciences departments. The second
group consisted of seven undergraduate university students of Mathematics
and Computer Science. All participants from both students groups had been
recruited via a word-of mouth or an announcement at Nicolaus Copernicus
University in Torun (Poland) whereby extra course credits were offered as well
as feedback about their performance on the study tasks. The last group
consisted of three experts from Computer Science Faculty at the same
University. Subjects from all three groups were tested individually. They were
asked about key aspects of visualization layout using a questionnaire.
The survey research was oriented towards substantial assessment of
interpretation of graphical representations by both individuals and groups. To
reach uniformity of output quantitative data, series of closed-ended questions
were constructed. Thematic categories labeled with letters A-K predefined an
exhaustive set of responses. Multiple-choice questions indicated the labels
(categories/classes). The authors added simple instructions at the beginning of
the questionnaire form to get respondents-amateurs acquainted with the topic.
The ten questions concerned four crucial characteristics of graphical
distribution of documents nodes related to the three-dimensional layout: (1)
dynamics of changes, (2) clustering degree, (3) closeness and (4) semantic
correlation between two contemporary infrastructure logistics, i.e. cloud
computing” and “grid computing”. The dynamics was analyzed through seeking
the most changeable patterns across three different years of publishing. First
quaternary questions were concerned to the dynamical or permanent
distribution. If a category was highly dynamical, it couldn’t be selected again for
the question about the non-changeable pattern. The next question:
overlapping/separate categories by seeking the most/least colours mixing
between groups?” - was related to significant clustering and even distribution
which are mutually exclusive. Some categories of documents indicated by
different colours were located close to each other; some of them were more
distant. The proper question was formulated: “What the categories are the most
distant from each other?”. Semantic distribution could be discovered by
tracking graphical pattern of articles by selected topic across a years. Cloud
computing, the most popular technology today, has evolved from networks
(sub)classes. Three last questions were designed to prove that dependence.
The final item in the questionnaire included all ideas and comments about the
implementation of interface.
4. Results
Evaluation of research based on comparison the true answers defined by
experts and respondents answers. Only the experts may properly comprehend
and interpret classification system evolution. Accuracy was defined as a simple
percentage quotient of correct to all answers. A wide choice of response items
(11 main categories plus negative option) in each question has determined the
survey results to a large extent. It was observed that respondents in each
group displayed two different cognitive styles during exploration of graphical
patterns. The first one related immediate reply and the second was
characterized by longer time to select questionnaire answering sequences. In
the last case, the choice was more reasonable. Those distinct analysis
approaches caused discrepancy in the results and complicated the estimation
Final accuracy varies in the range from 50% to 86% and strongly depends on
the nature of a question. Given response sequences were compared with the
ones accepted by the experts. The lowest accuracy is observed in two cases
related to clustering and overlapping of thematic categories. Finding clusters
was problematic as the respondents wrongly understood the concept and the
formulation of the question was unclear. That can be the reason of
comparatively low accuracy in relative responses. Semantic overlapping is
possible to be revealed through tracking the intensity of colour mixing.
According to suggestions of the experts, some colours combination create
difficulties in complex patterns perception. That problem concerns orange
(H._Information Systems Organization) and aqua (I._Computer Methodologies)
nodes distribution.
The best accuracy has been obtained for the following issues: dynamics
characteristics, even distribution as well as monitoring the cloud computing
patterns. It was noted that Computer Science and Humanities students
exposed different behaviour during interface analysis. Humanities students
focused on the practical application (information retrieval, Web browsers,
digitalization, library resources cataloguing) and knowledge domains (Library
and Information Science, computer science, robotics, education) of presented
visualization while computer science students were more interested in the
working principle of application and ignored appearing technical weaknesses of
interface like large delay time or lack of grid sphere.
5. Discussion and Conclusion
Obtained visualization maps might be used for methodological study of original
classification, its development, structure, dynamics as well as efficiency
regarding users. Classified documents form a complex pattern according to
their thematic similarity. The main assumption is that output graphical pattern
corresponds to the semantic structure of classification using relevant
topological space (sphere surface in current paper). Professionals who are
competent in both computer science and science of science issues are able to
comprehend output visualization maps and finally interpret them in the most
rational manner. The experts play such a role in the proposed experiment.
Users-amateurs are involved simultaneously in the process of visual reading.
The main objective is based on comparison of users-nonprofessionals
perception of maps with experts’ insight. The results could provide a basis for
quantitative evaluation of classification visualization approach.
Accuracy of responses does not fall below 50%. The best accuracy (86%)
relates to both dynamics and cloud/grid computing distribution analysis.
Clustering and overlapping of (sub)classes observations deal with great
discrepancy (lowest accuracy). Suitable questions are biased by insufficient
understanding and perception mechanism, for instance, scattered nodes of
isolated category were perceived as cluster. Besides, distance estimation
(proximity task) on sphere surface is problematic due to lack of texture,
coordinates and reference point (something like Greenwich). Experts have also
noticed the difficulty with colours differentiation on a black background.
Except those weaknesses, visualization interface is positively assessed by
users. Interaction and spherical configuration provides the convenient
exploration. By rotating a sphere, it is possible to see a graphical pattern
created by all categories and simultaneously to investigate the distribution of
nodes locally. That technique is widely used in Infoviz and it is called
focus+context (Osinska 2010, Young 1996).
This pilot study was a first step to answer the question how to prove that
visualization method and interface. Distinct groups of users had to investigate
classification visualization using interactive interface. They demonstrate distinct
cognitive styles according to their individual cognitive processes. The authors
plan to extend experiment with a large users group (more that one hundred).
Parallely all experts observations and suggestions must be taken into
consideration. Improvement of survey and interface includes:
- better clarification of formulation of such concepts like clustering,
overlapping, closeness;
- consideration of distinct cognitive styles of users;
- sphere rendering including grid, coordinates and reference points;
- textured sphere surface;
- better perceptible colours combination;
Proposed visualization method which provides a nonlinear exploration space
can be very useful for librarians, classifiers, information specialists and all
scientists from interdisciplinary research fields.
Borner, K. (2010). Atlas of Science, MA: MIT Press.
Börner, K. (2010) Extracting and Visualizing Semantic Structure in Retrieval Results for
Browsing. In: Proceedings of the fifth ACM conference on Digital Libraries, NY,
USA:ACM, 2010. Available at:
Boyack, K W. et al. (2005). Mapping the backbone of science. Scientometrics. Vol. 64,
no. 3, pp. 351-374. Available at:
Birger, H. (2013). User-based and Cognitive Approaches to Knowledge Organization: A
Theoretical Analysis of the Research Literature. Knowledge Organization. 40(1).
Carretié, L. et al. (2003). Cerebral patterns of attentional habituation to emotional visual
stimuli. Psychophysiology, 40, pp. 381-338.
Klavans, R.; Boyack, K. (2006). Quantitative Evaluation of Large Maps of Science.
Scientometrics 68 (3): pp. 475-499, 2006. Available at:
Krzywinski, M. et al. (2009). Circos: an information aesthetic for comparative genomics.
Genome Research, 19(9). Available at:
Osinska, V. (2010). Visual Analysis of Classification Scheme. Knowledge Organization,
Osinska, V.; Bala, P. (2010). New Methods for Visualization and Improvement of
Classification Schemes the case of computer science. Knowledge Organization,
Exhibit Purpose and Goals [online] (2007- ). Places@Spaces: Mapping Science.
Available at:
Szelag, E.; Dreszer, J.; Lewandowska, M.; Medygral, J.; Osinski, G.; Szymaszek, A.
(2010). Time and Cognition from the Aging Brain Perspective: Individual Differences.
Personality from biological, cognitive and social perspectives. Eliot Werner Publications
INC, 2010, pp. 331-364.
Thurston, W. (1997). Three-dimensional geometry and topology. Princeton
Mathematical Series, Vol.1, 35. Princeton University Press, NJ.
Ware, C. (2004) Information Visualization: Perception for Design. Morgan Kaufmann,
pp. 11, 188, 273.
Young, P. (1996). Three Dimensional Information. [online]. Department of Computer
Science. Available at:
About authors
Veslava Osinska is an assistant professor at the Institute of Information Science and Book
Studies, Nicolaus Copernicus University in Torun where she teaches information architecture,
information visualization, ICT and computer graphics. She has a degree in physics and holds a
PhD in library and information science. Her research interest is in modern information and
knowledge domain visualization with particular interest in applications of nonlinear properties to
information organization and streaming. Veslava is also a member of the Polish Chapter of
International Society of Knowledge Organization and the Polish Computer Science Society
Joanna Dreszer-Drogorob holds a PhD in psychology. She is an assistant in the Multimedia Lab
at the Fine Arts Department at the Nicolaus Copernicus University in Torun where she also teaches
cognitive psychology. Her research interests are in developmental aspects in time perception,
neural basis of human cognition, intelligence, cognitive abilities. She is also interested in the
dynamics in social psychology, such as dynamics in attitudes, feeling, and self-esteem.
Grzegorz Osinski is a computer scientist, neuroscientist and physicist. His current research is in
nonlinear dynamics systems in biomedical sciences where he applies numerical methods and
computer simulations for modeling behaviours of neural correlates in different dynamical states. His
other interests are in cognitive aspects of neural activity in perception and communication.
Michal Gawarkiewicz holds a degree in computer science from the Nicolaus Copernicus
University in Torun. His master’s thesis was on semantic memories in narrow domain using
machine-readable information. His current PhD studies are on database optimizations. Michal also
teaches programming classes and works in the area of data processing and mobile technologies.
... While reading a visualization map, cognitive perception structures are activated by means of analytical modules in a natural way, in contrast to the forced use of verbal narrative alignment. By analyzing the scheme presented above and building on the authors' own study (Osinska, Dreszer, Osinski, & Gawarkiewicz, 2013), it is possible to draw conclusions about the process of reading and analyzing science maps. The most striking is the searching for familiar problems, forms, or contents. ...
... When considering numerous examples of multivariate data mapping in the Places & Spaces collection, the authors relied on observations, assuming that the maps of knowledge and the maps of science could extend and make e-learning methodology more effective. Science maps, as an example of multidimensional presentation on a simple paper surface, constitute an interesting research material in terms of cognitive aspects in learning processes (Osinska, Dreszer, Osinski, & Gawarkiewicz, 2013). The main focus in making observation-based conclusions depends not only on age, profession, and interest areas but also on artistic sensibility of active followers. ...
... A typical e-learning system based on the appointed sequence of topics is perceived by students as boring, limited, and emotionally attached to a school learning model. The authors propose to apply a dynamical interactive interface, close to the natural spherical vision form that uses a fractal texture (Osinska et al., 2013). A pilot study has been carried out to explore the collection of scientific articles (Figure 15). ...
... Finally, the series of maps which depict the real changes of primary classification scheme indicate the most effective mapping technique in this case. In a parallel work the authors argue that a sphere surface is the most convenient shape for users' perception and cognition [6]. The authors focused on the way of how different groups of users, including experts can observe and identify the changes and evolution in a particular field of computer science. ...
... An original image mapped to the spherical retina also shows the characteristics of a spherical 3D structure [25]. It seems reasonable to use spherical data visualisation methods adapted to the natural abilities of the human vision [6]. There is a lot of open source software for graph and network analyses including wide interactive possibilities (Walrus, Gephi, Pajek etc.). ...
Full-text available
The authors compared three methods of mapping, considering the maps as a visual interface for the exploration of scientific articles related to computer science. Data were classified according to the original Computing Classification System (CCS) classification and co-categories were used for similarity metrics calculation. The authors’ approach based on MDS was enriched by algorithm mapping to spherical topology. Three other methods were based on VOS, VxOrd and SOM mapping techniques. Visualization of the classified collection was done for three different decades. Tracking the changes in visualization patterns, the authors sought the method that would reveal the real evolution of the CCS scheme, which is still being updated by the editorial board. Comparative analysis is based on qualitative methods. Changes in those properties over two decades were evaluated for the benefit of the authors’ method of mapping. The qualitative analysis shows clustering of proper categories and overlapping of other ones in the authors’ approach, which corresponds to the current changes in the classification scheme and computer science literature.
... Qualitative approach to compare visualization maps [10] shows the similar results: both MDS-sphere and SOM reveal consequence in dynamics changes, moreover VOS and VxOrd -inappropriate topology for data exploration [10,Appendix] How easy users can play with data and analyze their change -show the ergonomic properties of visualization interface. Homogeneous occupation of visual layout, edgeless, continuity in changes should feature good visualization [23]. Several papers described particularly this application from the end-users-analysts point of view [10], [24]. ...
Conference Paper
Full-text available
The paper contributes to the problem solving in semantic browsing and analysis of scientific articles. With reference to presented visual interface, four – the most popular methods of mapping including own approach - MDS with spherical topology, have been compared. For a comparison quantitative measures were applied which allowed to select the most appropriate mapping way with an accurate reflection of the dynamics of data. For the quantitative analysis the authors used machine learning and pattern recognition algorithms and described: clusterization degree, fractal dimension and lacunarity. Local density differences, clusterization, homogeneity, and gappiness were measured to show the most acceptable layout for an analysis, perception and exploration processes. Visual interface for analysis how computer science evolved through the two last decades is presented on website. Results of both quantitative and qualitative analysis have revealed good convergence.
... Qualitative approach to compare visualization maps [10] shows the similar results: both MDS-sphere and SOM reveal consequence in dynamics changes, moreover VOS and VxOrd -inappropriate topology for data exploration [10,Appendix] How easy users can play with data and analyze their change -show the ergonomic properties of visualization interface. Homogeneous occupation of visual layout, edgeless, continuity in changes should feature good visualization [23]. Several papers described particularly this application from the end-users-analysts point of view [10], [24]. ...
Conference Paper
Full-text available
Large-scale distributed digital library systems with aggregated metadata provide platforms for resource discovery and retrieval. For researchers, aggregated metadata offers a potential for big data analysis and exploration of digital knowledge growth. The paper reports the findings of the study that investigated the distribution of the date elements in the metadata aggregated in the Polish Federation of Digital Libraries and related it to the types of libraries. The purpose of this study was to address the gap in research about heterogeneous digital libraries and explore the dynamics of their growth. The authors included timeline characteristics of the development of Polish digital libraries and proposed a new dynamics parameter – resource release interval. They used histograms, which have been grouped according to the organizational and thematic criteria, developed for this study. All charts are characterized by two similar maximum points. Their shapes and ratio have been analysed by both statistical and visual methods. The shape of resource release interval charts revealed characteristic differences for libraries types. The proposed approach, based on time characteristics, is an important step in the development of systematic classification of digital libraries and digitizing institutions. It can be also considered as a new tool in monitoring the dynamics of digital knowledge growth.
Full-text available
Altmetrics, as a new scientometric measure, shows the scholarly impact of a scientist and her/his social activity on social networking platforms. We analysed communication on ResearchGate within the group formed by Nicolaus Copernicus University different domains researchers. They share interest in information visualisation and form local collaboration team. We considered two compositions of research sample: the original team and the nearest collaborators. The common attributes of RG users such as coauthorship, skills set, topics, and ascribed domains were analysed. Based on these units, we carry out the sociographs to reveal social structure dynamics. Domain analysis was performed by the use of the PCA – the method which is able to extract the most essential factors from the variables set. The results indicate a good coincidence for these two approaches: collaboration between researchers and their skills development can significantly change their major. We strived to show that RG is a space of interdisciplinary training in building and identifying their own field research based on full text. Thus RG is a proper tool for crossdisciplinary qualitative analysis. Altmetrics and visualisation are mutually substituted and complemented devices, perceived as new tools for team science study. This requires, nevertheless, not only the use of appropriate tools, but also collecting a large dataset on global scientific activity within multidisciplinary groups.
Full-text available
Visualization is currently used as a data analysis tool and considered a way of communicating knowledge and ideas in many areas of life such as science, education, medicine, marketing, and advertisement. The chapter contains complex interdisciplinary material and attempts to construct a general framework of visualization roles in learning. The structure of this content is presented on Figure 1 as a kind of mind map. Authors emphasize that visualization mechanisms are designed taking into account analytical and content-related potential, timeliness, online availability, and aesthetics. But classical (tabular) forms still remain dominant in information presentation. A good portion of the discussion is dedicated to the alternative solution – non-linear layout, such as network or fractals. Several visualization maps with specifically designed architecture are demonstrated as important elements of contemporary education. Authors consider the potential and implementation of these tools in e-learning platforms. Parallely, they underline the role of interdisciplinary collaboration in map making processes. Researchers in different fields can apply contemporary trends in visualization including natural shape perception, 3D representation problems, as well as the aspects of neuroaesthetics.
Full-text available
Umiejętność analizowania i wykorzystywania dużych zbiorów danych w praktyce dydaktycznej zaczyna odgrywać coraz większą rolę zarówno na polu metodyki nauczania jak i praktyki dydaktycznej. Współczesne trendy w nauce światowej jasno wskazują na coraz większą unifikację na polu nauk przyrodniczych. Wykorzystanie interaktywnych systemów sieciowych pozwala prześledzić zmianę struktury nauki współczesnej z uwzględnieniem wyników najnowszych badań.
Full-text available
Full-text available
In the 1970s and 1980s, forms of user-based and cognitive approaches to knowledge organization came to the forefront as part of the overall development in library and information science and in the broader society. The specific nature of user-based approaches is their basis in the empirical studies of users or the principle that users need to be involved in the construction of knowledge organization systems. It might seem obvious that user-friendly systems should be designed on user studies or user involvement, but extremely successful systems such as Apple’s iPhone, Dialog’s search system and Google’s PageRank are not based on the empirical studies of users. In knowledge organization, the Book House System is one example of a system based on user studies. In cognitive science the important WordNet database is claimed to be based on psychological research. This article considers such examples. The role of the user is often confused with the role of subjectivity. Knowledge organization systems cannot be objective and must therefore, by implication, be based on some kind of subjectivity. This subjectivity should, however, be derived from collective views in discourse communities rather than be derived from studies of individuals or from the study of abstract minds.
Full-text available
Generally, Computer Science (CS) classifications are inconsistent in taxonomy strategies. Itis necessary to develop CS taxonomy research to combine its historical perspective, its current knowledge and its predicted future trends - including all breakthroughs in information and communication technology. In this paper we have analyzed the ACM Computing Classification System (CCS) by means of visualization maps. The important achievement of current work is an effective visualization of classified documents from the ACM Digital Library. From the technical point of view, the innovation lies in the parallel use of analysis units: (sub)classes and keywords as well as a spherical 3D information surface. We have compared both the thematic and semantic maps of classified documents and results presented in Table 1. Furthermore, the proposed new method is used for content-related evaluation of the original scheme. Summing up: we improved an original ACM classification inthe Computer Science domain by means of visualization.
Full-text available
This paper proposes a novel methodology to visualize a classification scheme. It is demonstrated with the Association for Computing Machinery (ACM) Computing Classification System (CCS). The collection derived from the ACM digital library, containing 37,543 documents classified by CCS. The assigned classes, subject descriptors, and keywords were processed in a dataset to produce a graphical representation of the documents. The general conception is based on the similarity of co-classes (themes) proportional to the number of common publications. The final number of all possible classes and subclasses in the collection was 353 and therefore the similarity matrix of co-classes had the same dimension. A spherical surface was chosen as the target information space. Classes and documents' node locations on the sphere were obtained by means of Multidimensional Scaling coordinates. By representing the surface on a plane like a map projection, it is possible to analyze the visualization layout. The graphical patterns were organized in some colour clusters. For evaluation of given visualization maps, graphics filtering was applied. This proposed method can be very useful in interdisciplinary research fields. It allows for a great amount of heterogeneous information to be conveyed in a compact display, including topics, relationships among topics, frequency of occurrence, importance and changes of these properties over time.
Full-text available
Most designers know that yellow text presented against a blue background reads clearly and easily, but how many can explain why, and what really are the best ways to help others and ourselves clearly see key patterns in a bunch of data? This book explores the art and science of why we see objects the way we do. Based on the science of perception and vision, the author presents the key principles at work for a wide range of applications--resulting in visualization of improved clarity, utility, and persuasiveness. The book offers practical guidelines that can be applied by anyone: interaction designers, graphic designers of all kinds (including web designers), data miners, and financial analysts.
Conference Paper
Full-text available
The paper introduces an approach that organizes retrieval results semantically and displays them spatially for browsing. Latent Semantic Analysis as well as cluster techniques are applied for semantic data analysis. A modified Boltzman algorithm is used to layout documents in a two-dimensional space for interactive exploration. The approach was implemented to visualize retrieval results from two different databases: the Science Citation Index Expanded and theDido Image Bank.
Full-text available
This article describes recent improvements in mapping the world-wide scientific literature. Existing research is extended in three ways. First, a method for generating maps directly from the data on the relationships between hundreds of thousands of documents is presented. Second, quantitative techniques for evaluating these large maps of science are introduced. Third, these techniques are applied to data in order to evaluate eight different maps. The analyses suggest that accuracy can be increased by using a modified cosine measure of relatedness. Disciplinary bias can be significantly reduced and accuracy can be further increased by using much lower threshold levels. In short, much larger samples of papers can and should be used to generate more accurate maps of science.
Data Spaces) and implements a number ofalgorithms to structure data according to these key features.3.4.1. DistrictsDistricts are distinct areas within an environment which can be abstracted into a single entity. Suchareas must usually contain some form of commonality or character in order for them to be considered asa whole. Districts can be identifiable in a real world sense, for example by the nature of the buildingswithin them. An example of two districts could be a residential...