EXTENDED VISUALIZATION FOR A DIGITAL JOURNAL

Muhammad Salman Khan, Muhammad Tanvir Afzal, Narayanan Kulathuramaiyer

and Herman Maurer

Institute for Information Systems and Computer Media, Technical University of Graz,Graz, Austria

Keywords: Temporal, Geographical, Categorical, Bubble chart, Pie chart.

Abstract: Content analysis has been a tradition of many electronic and printed journals, in order to ensure quality and

the journal’s standing. Traditionally, researchers have tried to analyze patterns in scholarly publications

using normal tables and statistical charts. In this paper we present an interactive visualization system that

can help for a deeper analysis of different trends’ patterns hidden in scholarly publications of a digital

journal. We apply this technique to the Journal of Universal Computer Science (J.UCS). The proposed

visualization system is an easy to use web application, based on animated 2D bubble chart and pie chart to

handle geographical, temporal and large kinds of categorical data. The paper gives a brief overview of the

state of the art visualization techniques available to understand the knowledge structure of any given

academic discipline. The design and technical aspects of the proposed visualization tool and various

interesting results drawn from it have been discussed.

1 INTRODUCTION

In any academic discipline research publications

represent the knowledge structure of that discipline.

This knowledge structure reflects the history,

research trends, social structure of researchers,

networks of scholarly papers, experts, key papers,

contributions and collaborations of institutions and

regions. Much can be learned by analyzing the

research contributions in a journal or conferences of

any discipline about a given field of study (Taylor,

2001). This practice of analyzing publications has

been a tradition of many printed and electronic

journals. During the last few decades many studies

have been conducted to analyze the publications

patterns by length of articles, citations, affiliations

and geographical distribution of authors,

contributions in different research areas, trends of

research areas over time by analyzing publications

of one or more academic journals (Taylor, 2001),

(Hawkins, 2001). The benefits of such analysis are

enormous; first, it helps the administration of a

journal in increasing its quality by determining the

coverage and impact of the journal and the journal’s

standing, secondly it can be used to evaluate

individuals, organizations, groups and nations which

in turn may be used to inform the impact of

decisions and policies made for allocating resources

and funds, and thirdly it reduces the researchers’

menial efforts to conduct their surveys and shows a

broader picture for them to understand the field of

their interest (Boerner et al., 2003). In our previous

work (Khan et al., 2008) we explored the usage of

mash-ups, an emerging Web 2.0 technology to

strengthen the internal administration and providing

better facilities for the authors and readers of a

digital journal. The proposed system was a good tool

in highlighting the geographical coverage of

publications and editors in any particular research

area, bias groups of authors and editors in the review

process of articles, selection of special issues

according to the geographical policy of the journal,

novel navigational features and determining research

collaborators for authors and readers. But the system

was limited in answering number of issues such as:

how the papers, authors and institutions

contributions and research interests have changed

overall across the glob or in a particular location

over the period of time to help in promoting the

journal globally or at any particular location, in

finding which research area is becoming localized to

a specific community, which research area is

evolving or declining to assist in making decisions

regarding the call and acceptance of special issues

and acquiring reviewers accordingly, which research

area is becoming hot for researchers. In this paper

we introduce an interactive user-friendly web based

visualization technique that can help for a deeper

analysis of different trends’ patterns in scholarly

publications of a digital journal and its field of study

385

Khan M., Afzal M., Kulathuramaiyer N. and Maurer H.

EXTENDED VISUALIZATION FOR A DIGITAL JOURNAL.

DOI: 10.5220/0001754303850388

In Proceedings of the Fifth International Conference on Web Information Systems and Technologies (WEBIST 2009), page

ISBN: 978-989-8111-81-4

over time.

Traditionally, researchers have tried to analyze

patterns in scholarly publications using tables and

statistical charts. Interactive visualizations has been

used by (Ke et al., 2004), (In-SPIRE, 2004) to

realize different patterns such as citations network of

publications, number of papers over time and the

correlated research areas in the publications

published during the 8 years of InfoVis conferences.

In (Erten et al., 2004) authors have used 100,000

unique ACM computer science papers and analyzed

with the help of interactive node link graph the

evolution of different research areas and the

collaborative network of scientists in the field of

computer science. All the systems mentioned above

are a good tool in understanding either the network

of papers, authors and research areas or how the

research areas have emerged over the time but they

do not demonstrate the change of interest in

publications contributions and research areas across

different regions. NetLens (Kang et al., 2007)

applied a different approach where interactive bar

charts, list view and multiple coordinated windows

were used to analyze and compare trends over time

among different research areas, different countries,

important authors, papers and institutions for CHI

conferences. But the adopted visualization approach

is handling a fewer number of research areas

(categories) of the papers, which if increased can put

cognitive load to the users while comparing trends

of different research areas over the time. Moreover

the users can not compare the contributions from

different locations with each other over the period of

time (how different locations have progressed as

compared to others).

The proposed web based visualization tool

elaborated by this paper uses a simple animated 2D

bubble chart and pie chart for handling multivariate

(Geographical, Temporal and large Categorical)

data, and demonstrates deeper trends’ patterns over

time for the authors and administration of a journal

and the researchers of its field. We use J.UCS as an

example for our analysis. J.UCS is an open access,

high quality, peer reviewed electronic journal having

more than 1000 publications since 1994. The journal

covers all aspects of computer science discipline

(J.UCS, 2008).

2 TECHNICAL ASPECTS OF THE

VISUALIZATION TOOL

In order to make J.UCS a high quality peer reviewed

journal and maintaining its international standing,

the administration, authors and readers of J.UCS

needs to visualize the current status of the journal.

There is a need to visualize various factors such as

the rise or decline of research areas, the coverage or

impact of the journal, authors and institutions

participations and the trends of length of papers

across different research areas and locations.

2.1 J.UCS Data Extraction

The J.UCS document collections and their metadata

are stored and managed using the Hyperwave system

(Hyperwave, 2008). Publications in J.UCS have a

corresponding metadata (XML) file, one for each

paper that contains all the information about it such

as title, authors, institutions, volume, issue etc. In the

development of the visualization tool, metadata

about papers published was captured from the

Hyperwave server using Hyperwave APIs and stored

in a relational database.

2.2 Visualization Tool Design Choices

By keeping in view the initial requirements, the data

contains three dimensions as follows:

1. Temporal

2. Categorical (More than 400 three level ACM

categories and two additional categories i.e.

“Science and Technology of Learning” and

“Knowledge Management”)

3. Geographic

A detailed survey (highlighted in section 1) of

various available visualization tools that can fulfil

the requirements mentioned above has been

conducted. The motivation was to develop a user

friendly, easy to understand trend analyzer that

targets not only the experts but also general

academic users. An appropriate choice was to use

Gapminder (Gapminder, 2008), which can visualize

geographical trends over time in the form of

animated bubble charts. But the limitation of this

tool is that it does not cater to categorical data,

which in our case are ACM categories. On the basis

of Gapminder a visualization tool was implemented

that uses bubble chart to illustrate geographical

trends and a pie chart to depict details about

categories for any location.

3 EXPERIMENTAL RESULTS

In this section, some interesting results are presented

that can be obtained with this visualization tool. The

main interface of the visualization tool is shown in

Fig. 1. The user can select to view patterns in the

publications published in regular issues, special

issues or both.

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

386

Figure 1: Main Interface.

Figure 2: Distribution of publications for the level-1 categories.

The user can also select to view patterns in the

whole world as a single entity

or across countries and regions. The results can also

be filtered by selecting any topic or country from the

list of countries and topics. A temporal slider has

been provided for the users to scroll across different

years. The user can also play the slider for automatic

scroll across the years and can view the animated

moving bubbles and pie chart to reveal different

patterns. Each bubble on the chart represents a

country, region or the whole world based on the user

selection. The color and size of the bubbles

represents the location and number of publications

respectively. The axis of the animated bubble chart

contains various options in which the user might be

interested, such as number of institutes, number of

authors, number of papers, average length of papers,

and average number of authors per paper. The pie

chart represents the distribution of publications

across topics for any particular country, region or the

whole world. In order to facilitate an easy analysis

and to better understand the results, the publications

of J.UCS from 1994-2008 have been divided in three

groups; each spanning 5 years i.e. 1994-1998, 1999-

2003 and 2004-2008 (so far papers for the first six

months of year 2008 are available). In this paper we

are considering mainly regular issues instead of

special issues of J.UCS for our analysis because they

represent a clear picture about various trends. The

following sub-sections represent some interesting

results with regards to three different views.

3.1 World View

This view reflects all publications in J.UCS as a

single entity. The results demonstrate the evolution

of J.UCS with regards to the number of publications,

institutions and authors for the regular issues. The

total number of publications, authors and institutions

in J.UCS were 130, 206 and 200 respectively up to

1998. It is observed that there is a consistent decline

in publications for the time period 1999-2003 (90)

and 2004-2008 (84), whereas authors and

institutions first declined for the time period 1999-

2003 (140 authors, 133 institutions) and then started

to increase in 2004-2008 (177, 156). These statistics

also reflect the inclusion of new authors and

institutions in the journal instead of being occupied

by some groups of authors. Fig. 2 demonstrates the

distribution of publications across different research

areas. It is observed that the two top most research

areas are “Theory of Computation” (1994-1998: 27,

1999-2003: 32, 2004-2008: 22) and “Information

Systems” (1994-1998: 28, 1999-2003: 20, 2004-

2008: 28). The same view can also be used to

visualize research areas that have started to diminish

or grow. For example up to 1998 there was only one

publication in the research area “Computer

EXTENDED VISUALIZATION FOR A DIGITAL JOURNAL

387

Applications”. Then there is a sudden rise in

publications (11) from 1999-2003 and again a

decline in publications (2) from 2004-2008. A

similar phenomenon happened with “General

Literature”. The user has the choice to view the

similar trends in the sub-categories of any research

area by clicking any top level category on the

animated pie chart. It has been observed that

“Software Engineering” (a sub- category of

“Software”) was not a dominant research area when

compared to “Programming Languages” up to 1998,

but it started to evolve from 1999-2008 and is now

the most dominant research area in its category. The

emergence of “Software Engineering” can be further

validated by considering both special and regular

submissions. Such analysis of research areas is

necessary as it gives an overview to the new

researchers about the emergent or hot research areas

of their field. Moreover it helps the administration of

the journal to acquire reviewers for each research

area accordingly, for the call and acceptance of

special issues and to ensure that the coverage of the

journal in each research area remains global instead

to a particular locality.

3.2 Regional View

This view demonstrates the distribution of

publications across different regions. The results

show that European countries remain as the main

source of publications in the journal for all the time

periods but there is consistent decline of publications

(1994-1998: 92, 1999-2003: 67, 2004-2008: 50)

with the passage of time from Europe. Further

analysis revealed that the Asian countries including

China, India, Pakistan, Japan, South Korea and

Russia are consistently contributing more

publications (1994-1998: 6, 1999-2003: 6, 2004-

2008: 10) in the journal.

3.3 Countries View

This view further provides more insight into

publications patterns. It enables the users to

understand the participation of each country in the

journal, when a country started to contribute, when it

stopped to contribute, who is contributing more or

less in a research area. The results reflect that most

of the publications in J.UCS have been contributed

by Austria (1994-1998: 24, 1999-2003: 16, 2004-

2008: 12) followed by Germany. Further analysis

revealed that most of the authors (1994-1998: 31,

1999-2003: 27, 2004-2008: 13) and institutions

(1994-1998: 31, 1999-2003: 21, 2004-2008: 11)

participations are from Germany. Interestingly

Finland and New Zealand were contributing

frequently in the journal for the first two time

periods but each of them has contributed only one

publication from 2003 to 2008. The administration

of the journal in this case can take action to

encourage researchers in these locations to submit

their papers in the journal.

4 CONCLUSIONS AND FUTURE

WORK

A simple approach using 2D animated charts has

been demonstrated that handles geographic,

temporal and large kinds of categorical data to

realize hidden trends in scholarly publications of a

digital journal. Our experimentations conclude that

the proposed web based visualization system is a

powerful tool in determining the impact, coverage

and the status of the journal at deeper level.

Moreover it also draws a broader picture for the

researchers about the field of their interest.

In future, we have to further investigate the

proposed visualization tool by usability testing for

further improvements. We would like to enlarge our

data by including papers from publicly available

databases (DBLP, CiteSeer).

REFERENCES

Boerner, K., Chen, C. and Bayak, K. W., 2003.

Visualizing Knowledge Domains, Blaise Cronin (Ed.)

Annual Review of Information Science and

Technology, Inc/American Society for Information

Science and Technology, chapter 5, pp. 179-255.

J.UCS, 2008. http://www.jucs.org

Ke, W., Boerner, K. and Vishwanath L., 2004. Major

Information Visualization Authors, Papers and Topics

in the ACM Library. In InfoVis’04, IEEE Symposium

on Information Visualization.

In-SPIRE, 2004.http://in-spire.pnl.gov/

Erten, C., Harding, P. J., Kobourov, S. G. and Wampler,

K., 2004. Exploring the computing literature using

temporal graph visualization. (Tech. Rep. TR0304).

Kang, H., Plaisant, C., Lee, B. and Bederson, B. B., 2007.

NetLens: iterative exploration of content-actor

network data, IV, 6, 1, pp. 18-31.

Gapminder, 2008. http://www.gapminder.org/world.

Hyperwave, 2008. http://www.hyperwave.com/e/.

Khan, M. S., Kulathuramayer, N., Maurer, H, 2008.

Applications of Mash-ups for a Digital Journal. Journal of

Universal Computer Science, 14, 10, pp. 1695-1716.

Taylor, E. W., 2001. Adult Education Quarterly From

1989 To 1999: A Content Analysis of All

Submissions. AEQ, 51, 4, pp. 322-340.

Hawkins, D. T., 2001. Bibliometrics of electronic journals

in information science. IR, 7, 1.

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

388