Facilitating the Handling of Documents

through Tree-Map Visualization

Elis Hernandes

, Erika Hohn

, José Carlos Maldonado

and Sandra Fabbri

Software Engineering Research Lab (LaPES), Universidade Federal de São Carlos, São Carlos, Brazil

Institute of Mathematics and Computer Science (ICMC), Universidade de São Paulo, São Carlos, Brazil

Keywords: Tree-Map, Information Visualization Tool, Hyperlink Function, Data Analysis, Structured Reading, Visual

Analysis.

Abstract: Applying visualization to manipulate a large volume of data constitutes a current demand of professionals

and researchers. Visualization stimulates human perception and increases the understanding capacity. In

literature, there are many studies on the use of visualization in different areas, as well as tools that make its

use feasible. Aiming to facilitate user’s comprehension, the objective of this paper is to explore, in the tree-

map visualization, the combined use of the following functions: search, edit, data regrouping and hyperlink.

This proposal was derived from two real problems: 1) to standardize a great volume of educational

assessment questionnaires that could not be processed automatically; and 2) to analyze the dependencies

among the TMMi (Test Maturity Model Integration) process areas, having simultaneous access to the text of

all the ones related to a specific issue. The functions were implemented in the SeEd-Visual tool (Search and

Edition based on Visualization), and organized in a process that aims to enhance user’s comprehension

through a visual metaphor. The process and the tool were applied to the real problems and allowed their

solution. The use of visualization enabled the solution of the real problems since it would be hard if done

manually.

1 INTRODUCTION

Nowadays, visualization has been used for different

proposals as scientific data analysis, knowledge

discovery, data mining, etc. (Rübel et al., 2010) (Silva

et al., 2011). According to Card, Mackinlay and

Shneiderman (1999), visualization focuses on

mapping data into a visual form, aiming at supporting

human interaction usually through a tool for visual

sense making.

Burley (2010) emphasizes that information

visualization is a valuable tool for knowledge

integration activities. This fact has been increasing the

visualization application area which is driven by a

mix of practical needs and research interest.

In this paper, we show the use of visualization to

make easier the solution of two real problems. In both

cases, to solve the problem it was needed to handle a

large document. In one case the automatic handled of

the document was not possible and, in the other one,

the manual handle of a big document was infeasible.

Once identified that visualization could simplify

the problems solution, the development of SeEd-

Visual (Search and Edition based on Visualization)

tool was started. This tool uses tree-map visualization

combined with the following functions: search,

edition, data regrouping, and hyperlink navigation.

The use of these functions allows finding a specific

data highlighting it in the metaphor; modifying data

generating real-time regrouping in the metaphor; and

opening external documents associated with a specific

data found through the search function. The

intervention in the visualization associated to the

regroup and hyperlink functions facilitates the

comprehension and the analysis of large documents.

The rest of this paper is organized as follows: in

Section 2, some researches related to the work

presented here; in Section 3, it is explained how the

functions search, edit, regroup and hyperlink impact

the visualization process and also how the SeEd-

Visual tool works. Section 4 presents the two practical

needs that motivated the development of the SeEd-

Visual tool and how these problems were solved by

using this tool; and finally, Section 5 presents the

conclusions and further work.

Hernandes E., Hohn E., Maldonado J. and Fabbri S..

Facilitating the Handling of Documents through Tree-Map Visualization.

DOI: 10.5220/0004003300980104

In Proceedings of the 14th International Conference on Enterprise Information Systems (ICEIS-2012), pages 98-104

ISBN: 978-989-8565-11-2

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

2 RELATED WORK

In literature, there are many researches that explore

the use of visualization to support the execution of

different tasks and to facilitate the resolution of

different problems. In this section, some researches

undertaken in the context of document handling and

of characterizing functions provided by visualization

tools are commented.

Deep Diffs (Shannon et al., 2010) is a tool that

allows the analysis of document evolution through the

view of all the older versions. Parts of the text that

have been changed over multiple successive versions

are highlighted by the tool, drawing attention to

passages that are new, potentially unpolished or

contentious. This tool can be applied in a post-hoc

way to any text document stored in a version control

system integrated into a wiki, a source-code revision

system or a content-management system (CMS).

Deep Diffs helps the writing and the editing processes

by assisting memory and encouraging the analysis of

collaboratively-authored documents, through a text

editor and viewed by the tool with application of

different colors.

Another tool is the ResultMaps (Clarkson et al.,

2009) that helps the analysis of digital library lists

which are organized through. tree-map visualization.

Once the user has applied a query, the result is

showed as a list of items, which are highlighted in the

tree-map visualization. From this point, in case the

user navigates on the result list, data on these items

are presented in the visualization. In this case, the

interaction between the user and the visualization

happens through brushing.

OpinionSeer (Wu, 2010) is a tool developed to

facilitate the analysis of customers’ opinion on a

product or service. These opinions have been

collected online, which encourages their post and

generates a large volume of data. Aiming at

facilitating the analysis, the tool provides a visual

presentation of these data through scatterplots and

radial visualization.

The family of tools called PEx – Projection

Explorer (Paulovich et al., 2007) has been applied to

analyze and to verify the relationship among many

documents, which makes this error prone task easier.

PEx is an example of tool that provides the analysis of

different types of textual documents. PEx-Image

(Eler, 2009) allows the analysis of an image set and

PEx-Web (Paulovich, 2008) allows the analysis of

data recovered by web searches. The three tools allow

the user to open the visualized documents (image,

PDF documents, text and web pages).

The TreeMap tool (HCIL, 2003) implements the

tree-map technique (Johnson, Shneiderman, 1991). To

promote an increasing knowledge on data, the Treemap

allows hierarchies, filters construction and color

exchange. However, even though the tool makes

different ways of modifying the visualization available,

it does not provide search and edit functions.

In summary, visualization tools have given

support to different tasks. There are tools that provide

different types of visualization and one of their most

common characteristics is an appropriated interaction

type. Although some of the tools deal with documents

– text, images and metadata –, none of them provides

functions like search and edit, as proposed in the tool

presented here.

3 VISUALIZATION PROCESS

Motivated by the real problem of analyzing and

standardizing a large volume of data, a visualization-

based process and a the SeEd-Visual tool, that

supports this process, have been defined. Other

versions of the tool are been used to help the

resolution of similar problems.

The first problem that triggered the definition of

the process and the initial version of the tool is:

 The context: (i) a set of questionnaires that were

applied periodically, during twenty years; (ii) each

questionnaire was composed of about seventy

questions; (iii) each year a questionnaire was

applied, the writing of a question previously used

was not necessarily the same; besides, new

questions could be inserted and old questions could

be removed;

 The problem: determining which questions were

applied in each year;

 Background: (i) an automatic comparison could

not be applied since the writing of many questions

were different and; in addition, a typo could also

interfere; (ii) handling the questionnaires by hand

was not feasible; (iii) making a decision in case

syntactically different questions were semantically

equal implies human analysis; (iv) visualizing

simultaneously, all questions that could be

semantically equal, seemed the best option.

Thus, considering the foregoing explanation,

visualization seems to be an alternative to aid the

problem solution. As seeing all the questions

simultaneously would facilitate the analysis, the tree-

map technique (Johnson and Shneiderman, 1991) was

defined as the starting point. This algorithm uses one

hundred percent of the screen and the color and the

size of the boxes are appropriated for large volume of

data. Based on this decision, each box would

FacilitatingtheHandlingofDocumentsthroughTree-MapVisualization

represent a question and the box color would

represent the year when the question was applied. In

addition, identical questions would determine the

grouping of the boxes. The advantage of using

treemap instead of list, for example, is the possibility

of attributing different representations, like color and

size, to distinguish different data.

However, these decisions were not enough.

Aiming at solving the problem, it would be necessary

to identify questions that were semantically

equivalent – then, the need of the search function; and

to edit the writing of these questions, aiming at

concluding the analysis – then, the need of the edit

function. Once any edition is made and the semantic

equal questions are changed into syntactically equal

questions, this information immediately impacts the

visualization by regrouping the questions.

Therefore, the search and the edit functions

promote a real time change in the visualization, which

facilitates the handling of these data. According to

Kan et al (2010), it is important that the user interacts

with the visualization, since it makes the information

easily understood. In Figure 1 Card’s process (Card et

al., 1999) was adapted to represent our process.

As shown in Figure 1, the search function impacts

only the visual metaphor. In the SeEd-Visual tool, the

usage of this function corresponds to the definition of

a keyword, which highlights the questions that have

that keyword. On the other hand, the edit function

impacts the tabular data, which impact the result of

the tree-map algorithm application, which, by its turn,

generates a new visualization. In the SeEd-Visual, the

use of this function corresponds to the selection of a

question and to the edition of its writing.

After the definition of this process and the

implementation of the corresponding tool version, the

following problem was raised:

 The context: the need of dealing with a large

document – the TMMi reference model. This

model is a guide for test process improvement

(TMMi Foundation, 2011).

 The problem: identifying dependencies among the

practices.

 Background: (i) as the available document of the

TMMi reference model is a PDF file, the search

function of PDF readers locates an occurrence of a

keyword at a time; (ii) handling the TMMi

document manually was not feasible; (iii) making

the decision on practices dependencies implies

human analysis.; (iv) seeing all the practices that

cover the same issue simultaneously seemed the

best option to identify dependencies.

By analyzing both problems, it is possible to observe

similarities between them. Again, the visualization

and the search function, as described here, could aid

to solve the problem. In this case, the edit function

was not necessary since the original document would

not be modified. However, accessing the whole text

of the TMMi was essential to decide on the

dependencies. Thus, the initial version was updated to

make the hyperlink function available. The

hierarchical structure of the PAs is represented by

nested boxes and all practices are represented on the

screen. The practice name is transformed into a

hyperlink, where the text of this practice is accessed.

As shown in Figure 1, the hyperlink function does

not impact the steps of the visualization-process. In

the SeEd-Visual, the use of this function corresponds

to clicking on the practice name and, then, the whole

text of this practice is open on the screen. The user

can open as many practices as needed.

The SeEd-Visual has been implemented in Java

by using the IDE NetBeans 6.7 (NetBeans, 2001).

The visual metaphor of the tool uses the Tree-map

technique (Johnson and Shneiderman, 1991) from the

Prefuse library (Prefuse, 2012).

It is important to observe that although there are

some tools that use the tree-map technique (Clarkson

et al., 2009); (Newsmap, 2010), none of them provide

the edit and the hyperlink functions as presented here.

These functions, although simple, have been

implemented in the SeEd-Visual since they are useful

to handle large volume of data or large documents. In

the next section, two case studies that exploring the

functions will be presented.

Figure 1: The process supported by the SeEd-Visual tool

(adapted from Card, Mackinlay, Shneiderman, 1998).

4 CASE STUDIES

In this section the problems that motivated the process

definition and the current version of the SeEd-Visual

tool are presented. The first problem – the one related

to the questionnaires – is presented as Case Study 1 and

the second – related to the TMMi – is presented as

Case Study 2.

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

100

4.1 Case Study 1: Questionnaires

Analysis

One of the basic instruments for the evaluation,

planning and support of the decision making on

public policy establishment is the production of agile

and accurate data. Therefore, the amount of data

should be large enough to allow this kind of analysis.

The Brazilian Institute INEP has recorded data for

twenty years now, aiming to characterize the

Brazilian educational sector. These data have been

collected through the national evaluation system,

which involves different types of evaluations,

generating a rich set of data – each evaluation is

composed of approximately eleven questionnaires

with about seventy questions each. In spite of the

amount of data, they are not standardized and are

distributed in different databases.

Mainly because some questions were written in a

different way, it was not possible to use an automatic

tool to identify when they presented the same

meaning. Therefore, this identification requires a

human decision. However, as the amount of data was

large, some computational support was necessary

since the task would be time consuming and error

prone if done just manually.

Hence, the visualization-based process was

defined and used to make this task easier. Certainly,

in case of having a visual scenario of the questions

status − groups of identical questions, questions that

were used only once, etc. – this information could

facilitate the questions analysis and standardization.

Aiming at facilitating the explanation of this case

study the data of SAEB (Basic Education Evaluation

System), which is the system evaluation for school

students (from 6 to 15 ages), were used. This

evaluation has been applied every two years and its

data corresponded to 66 questionnaires and 5.711

questions. The figures of this section represent a set of

questionnaires that correspond to the discipline

“Portuguese”.

Figure 2 shows how the process of Figure 1 was

applied in this case study. In summary, the

questionnaires must be prepared generating a txt file,

which is uploaded into the SeEd-Visual tool. The tool

transforms the txt file into a visual metaphor, where

each question is represented by one box of the tree-

map visualization. After that, the questions are

processed iteratively through the search and the edit

functions. Every time a question is edited, the tool

reorganizes the questions (boxes) automatically by

providing a new visual metaphor. Below, each step of

the instantiated process of Figure 2 is detailed.

1) Prepare raw data to be uploaded into the SeEd-

Visual tool= join all questionnaires in a single txt

file.

2) Upload the .txt file into the SeEd-Visual tool=The

data are visualized as shown in Figure 3, where:

a) each box corresponds to a question;

b) each box is coloured according to the year when

the question was applied;

c) white boxes that group a set of colored boxes

correspond to a question that was applied by

using the same syntax (see the top left corner);

d) isolated boxes correspond to questions that were

not applied more than once (not with the same

syntax) (see the bottom right corner);

e) the information presented when clicking on a

box contains: the question, the database attribute

name, the attribute type, the year when the

question was applied and the questionnaire

name.

Figure 2: Flowchart of the visualization-based process of

the questionnaire analysis.

3) Standardize the data= if there is any non faded

box, the user should use the search function

aiming to analyze the highlighted questions and,

if necessary, edit them. The following actions

must be executed:

a) Click on a non faded box and identify a

keyword;

b) Type the keyword in the search field.

Immediately, all the boxes that contain such

keyword will become gray, as indicated by the

circles in Figure 3;

c) Analyze the questions of the gray boxes and

decide if they have the same meaning;

d) Standardize the text of these questions (using the

edit function), by choosing the most

appropriated syntax. The SeEd-Visual tool

regroups the questions (boxes), automatically,

and fades their colors to indicate that these

questions have been analyzed;

e) Repeat actions (c) and (d), until the syntax of all

gray boxes are analyzed;

f) Go back to action (a) until all the boxes are

faded.

After the execution of step 3, the data represented

in Figure 3 generates the visualization of Figure 5(a).

FacilitatingtheHandlingofDocumentsthroughTree-MapVisualization

101

Observe that in Figure 5(a) there are fewer isolated

boxes and more questions that have been grouped.

The isolated boxes correspond to questions that that

have been applied only once.

Figure 3: The questions highlighted contain the same

keyword.

Figure 5(b) shows the impact of applying the

visualization-based process on the SAEB evaluation

data. For example, the numbers related to the

“Português” questionnaire, considered in this study,

are highlighted in this figure. After the usage of the

functions search and edit combined with the tree-map

visualization, the questions that were applied to all

questionnaires increased from 4 to 24 and the

questions that were applied in only one questionnaire

decreased from 94 to 25.

Observe a similar situation occurred for the

questionnaires of the other disciplines. In other words,

the application of this visualization-based process

enhanced the quality and the usefulness of the data,

since a more accurate analysis can be done and the

amount of common data increases.

4.2 Case Study 2: Structured Reading

After the development of the first version of the

SeEd-Visual tool, another problem has arisen: the

need to establish dependencies among the TMMi

practices (TMMi Foundation, 2011). This model is

reference model structured in Levels, Process Areas

(PAs), Goals and Practices. To solve the problem, it

was necessary to compare all practices, in pairs,

aiming to identify if there were dependencies between

them and, if so, determine their type.

Aiming to execute this comparison it was

necessary to identify the practices that addressed a

specific subject, ie, a keyword. Then, to analyze

possible dependencies, it was necessary to read the

practices description.

Considering the availability of the SeEd-Visual

tool, it was adapted to aid in the solution of this

problem. As the TMMi document has a hierarchically

organized text, the tree-map visualization was ideal to

present it and to highlight all the practices that

addressed a subject, through the application of the

search function. Thus, the major modification on the

tool was related to the structured reading, which was

necessary to handle the information appropriately.

This modification allowed the creation of a hyperlink

between an item of the TMMi structure and the

respective text in the TMMi document, which was

transformed into a HTML file.

In this case study the visualization-based process

presented in Figure 1 was instantiated in the flowchart

of Figure 4.

Figure 4: Flowchart of the visualization-based process of

the questionnaire analysis.

First, two files are generated from the TMMi

document. One of them is a txt file, which contains

the structure of the reference model (Process Area,

Goals, Practices and Practices Descriptions), and is

uploaded into the SeEd-Visual tool. The other one is a

HTML file that is used to access parts of the TMMi

document, depending on the item of the reference

model that was selected by the user. The tool

transforms the txt file into a visual metaphor, where

each practice is represented by a box of the tree-map

visualization. After that, the search function can be

used to identify the TMMi practices that address the

same subject, and then can be interdependent.

Below, each step of the process for structured

reading, shown in Figure 6, is detailed.

1) Prepare the TMMi information to be uploaded

into SeEd-Visual tool= create a .txt file that contains

the TMMi structure: Process area, Goal and a short

description of the Specific Practice.

2) Transform the TMMi file into a HTML= create a

HTML file correspondent to the TMMi document,

containing hyperlinks in the titles of the structures

(process areas, goals and practices). The hyperlinks

should be equal to those contained in the .txt file

uploaded into the SeEd-Visual tool.

3) Upload the .txt file into the SeEd-Visual tool=

the data are visualized as showed in Figure 6(a),

where:

a) each colored box corresponds to a practice, but

unlike case study 1, the colors do not represent

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

102

a special piece of information (A);

b) white boxes grouping a set of colored boxes

correspond to a goal (B);

c) the outer level of grouping, i.e., a set of goals,

correspond to a process area (C);

d) the information presented at the bottom side,

when clicking on a box contains: the practice,

its goal and process area, beyond the practice

descriptive text (D).

4) Upload the .txt file into the SeEd-Visual tool=

the data are visualized as showed in Figure 6(a),

where:

a) each colored box corresponds to a practice, but

unlike case study 1, the colors do not represent

a special piece of information (A);

b) white boxes grouping a set of colored boxes

correspond to a goal (B);

c) the outer level of grouping, i.e., a set of goals,

correspond to a process area (C);

d) the information presented at the bottom side,

when clicking on a box contains: the practice,

its goal and process area, beyond the practice

descriptive text (D).

Aiming to establish the dependencies, the user

must provide the subject for which he wants to find

the related practices. This is done through the search

function, as in case study 1. As soon as he provides a

keyword for that subject, all the boxes that contain

such keyword will immediately become gray, as

indicated in Figure 6(b). This visual result also

allows realizing the impact of that subject on the

TMMi reference model as a whole. As an example,

in Figure 6(b), the keyword “criterion” was provided

to the search function. The gray boxes indicate that

this word appears in three different PAs.

After the identification of the PAs that address the

desired subject, the user can access the text of the

TMMi practices to analyze if they present depen-

dencies, i.e, in case the application of a practice

depends on the other one. Each double-click on a gray

box opens a HTML file (that conteins the TMMi prac-

tice), which allows displaying the practices in parallel,

enabling the comparison in pairs (Figure 6 (b)).

The structured reading herein presented makes

the handling of TMMi document easier. In this case

the visualization provides a better comprehension of

the document and facilitates its analysis. In addition,

(a)

(b)

Figure 5: (a) The same data presented in Figure 3 after the analysis and standardization. (b) Numbers on the SAEB

metadata before and after the analysis supported by the visualization-based process.

(a)

(b)

Figure 6: a) TMMi visualized in SeEd-Visual tool. b) The result of the search function applied with the keyword “criterion”

(in gray) and the text of two practices selected through hyperlinks.

FacilitatingtheHandlingofDocumentsthroughTree-MapVisualization

103

this kind of reading can be applied to different

domains, for example, laws, standards and manuals.

5 CONCLUSIONS

This paper presented the use of tree-map visualization

combined with the functions search, edition, data

regrouping, and hyperlink navigation as a practical

solution for data analysis.The use of these functions in

a coordinated way facilitates the interaction of the user

with the tool, enhancing the human cognition capacity.

This proposal was derived from the need of

analyzing a large amount of data, but this analysis

would be easier in case all the data were visualized at

the same time. Thus, aiming to take advantage of the

full screen, the tree-map visualization has been chosen.

Besides, the data (a set of questions) must be corrected

(ie, edited), triggering immediate regrouping in the

visual metaphor. The aim was grouping the data as

much as possible. In addition, to identify which data

must be corrected, the search option should also be

available. Hence, this initial proposal – tree-map,

search, edition and data regrouping – was implemented

in the tool named SeEd-Visual, which has been the

basis to solve similar problems.

After this first problem, it was needed to establish a

practical way to handle the large document of the

TMMi reference model. It was necessary to read the

description of its practices with the aim of deciding if

there was any dependency among the ones that were

related to a specific issue. Again, a whole overview of

the document structure requires the tree-map

visualization. The search function solves the problem

of identifying which practices should be read. But the

reading still requires the access of the TMMi practice

full text. Then, the hyperlink function was added to the

SeEd-Visual, allowing a structured reading,

simultaneously, fragments of this large document.

In summary, the tree-map technique was a good

choice to provide a view of the whole data and the

functions search, edit, regrouping and hyperlink made

the interaction between the user and the visualization

easier. As further work, the SeEd-Visual will be used

as the basis for the development of a free tool that can

be used in the context of qualitative analysis to support

the coding technique application.

ACKNOWLEDGEMENTS

The authors wish to thank the Brazilian funding

agencies CNPq, CAPES and FAPESP, the INEP

institute, the Observatório da Educação project and

INCT-SEC (National Institute of Science and

Technology - Critical Embedded Systems - Brazil),

processes 573963/2008-9 and 08/57870-9.

REFERENCES

Burley, D. 2010. Information visualization as a knowledge

integration tool. Jounal of Knowledge Management

Practice, 11(4).

Card, S., Mackinlay, J. and Shneiderman, B., 1998.

Readings in Information Visualization: Using Vision to

Think. Morgan Kaufmann; San Francisco, 1999.

Clarkson, E, Desai, K. and Foley, J., 2009. ResultMaps:

Visualization for Search Interfaces. IEEE Trans.

Visualization and Computer Graphics, 15 (1).

Eler, D. et al., 2009. Visual analysis of image collections.

The Visual Computer, 25 (10), pp. 923-937.

Johnson, B., Shneiderman, B. 1991. Tree-maps: a space-

filling approach to the visualization of hierarchical

information structures. In. Conf. Visualization, IEEE

Press.

Kan, Z., Hu, C., Wang, Z. Wang, G. and Huang, X., 2010.

NetVis: A network security management visualization

tool based on treemap. In. Inter. Conf. Advanced

Computer Control, IEEE Press.

NetBeans, 2012 - Java IDE (Integrated Development

Environment) [online] Available at:

<www.netbeans.org> [Accessed 20 January 2012].

Paulovich, F., Oliveira, M., Minghim, R., 2007. The

Projection Explorer: a flexible tool for projection-based

multidimensional visualization. In. BR Symp. Computer

Graphics Image Processing, IEEE Computer Society.

Paulovich, F. et al., 2008. PEx-WEB: content-based

visualization of web search results. In. 12th Inter. Conf.

Information Visualisation, Elsevier.

Prefuse, 2012. Visualization toolkit [online]. Available at:

<http://prefuse.org > [Accessed 20 January 2012].

Shannon, R., Quigley, A. and Nixon, P., 2010. Deep Diffs:

visually exploring the history of a document. In Inter.

Conf. Advanced Visual Interfaces, ACM New York.

Rübel, O., et. al., 2010. Coupling visualization and data

analysis for knowledge discovery from multi-

dimensional scientific data. Procedia Computer Science,

1(1), pp. 1757-1764.

Silva, S., Santos, B. S. and Madeira, J. 2011. Using color in

visualization: A survey. Computers & Graphics, 35 (2),

pp. 320-333.

TMMi Foundation, 2009. The Test Maturity Model

Integrated - TMMi Reference Model [online] Available

at:<http://www.tmmifoundation.org/html/tmmiref.html

> [Accessed 20 January 2012].

HCIL, 2003. Human Computer Interaction Lab, University

of Maryland [online] Available at: <http://www.cs.umd.

edu/hcil/treemap> [Accessed 20 January 2012].

Wu, Y. et al., 2010. OpinionSeer: interactive visualization

of hotel customer feedback. IEEE Trans. Visualization

and Computer Graphics, 16(6), pp. 1109-1118.

Newsmap, 2010. [online] Available at:

<http://newsmap.jp> [Accessed 20 January 2012].

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

104