OntoDIVE: An Ontology for Representing Data Science Initiatives upon

Big Data Technologies

Vitor Afonso Pinto

1 a

and Fernando Silva Parreiras

2 b

Technology Department, Operational Technology for Mine, Plant and Expedition, Vale Mozambique, Tete, Mozambique

Laboratory for Advanced Information Systems, FUMEC University, Rua do Cobre, Belo Horizonte, Brazil

Keywords:

Ontology, Data Science, Big Data.

Abstract:

Intending to be more and more data-driven, companies are leveraging data science upon big data initiatives.

However, in order to reach a better cost-beneﬁt, it is important for companies to understand all aspects involved

in such initiative. The main goal of this research is to provide an ontology that allows to accurately describe

data science upon big data. The following research question was addressed: ”How can we represent a Initiative

of data science upon big data?” To answer this question, we followed Knowledge Meta Processes guidelines

from Ontology Engineering Methodology to build an artifact capable of explaining aspects involved in such

initiatives. As a result, this study presents OntoDIVE, an ontology to explain interactions between people,

processes and technologies in a data science initiative upon big data This study contributes to leverage data

science upon big data initiatives, integrating people, processes and technologies. It conﬁrms interdisciplinary

nature of data science initiatives and enables organizations to draw parallels between data science results for a

particular domain to their own domain. It also helps organizations to choose both frameworks and technologies

based on their technical decision only.

1 INTRODUCTION

Data is constantly created, and at an ever-increasing

rate. Mobile phones, social media, imaging technolo-

gies and several other examples create new data which

must be stored somewhere for some purpose (Dietrich

et al., 2015). IT technologies have made all devices,

equipment, and systems in automation domain intel-

ligent, communicable, and integrated from the ﬁeld

level to the operation level for seamless data ﬂow in

both directions. Thus, there are multiple technologies

for data acquisition, transmission, storage, modeling

and so on. Organizations know that studies that were

difﬁcult to conduct in the past time due to data avail-

ability can now be carried out (Liu et al., 2016). Or-

ganizations are also aware that the timely analysis and

monitoring of business processes are essential to iden-

tify non–compliant situations and react immediately

to those inconsistencies (Vera-Baquero et al., 2016).

Nevertheless, even with all the progress that has

been made, companies are still struggling with how to

capture insights that are not obvious. It is a problem

of how to discover meaningful relationships (Hurwitz

https://orcid.org/0000-0002-2731-0952

https://orcid.org/0000-0002-9832-1501

et al., 2015). In a general way, organizations have dif-

ﬁculties to leverage big data initiatives as they may

not know what exactly is involved in such initiatives.

Although this concept has been implemented by many

parties, there exists a number of misconceptions re-

lated to the concept from the aspect of understanding

and implementation of a project like this (Abdullah

et al., 2017). The lack of concepts and an increas-

ing list of new technologies creates a fuzzy environ-

ment where organizations do not know what they ex-

actly need to do and on the other hand consultants,

technology developers, standard publishers and re-

searchers do not know how to help organizations to

achieve their goals. This condition limits the usage

of advanced analytics tools, preventing the capture of

potential beneﬁts.

Main purpose of this study is to describe data sci-

ence initiatives upon big data by using an ontologi-

cal approach. Ontologies play a fundamental role in

bridging computing and human understanding (Par-

reiras, 2011). They have been used in several ﬁelds

as an engineering artifact with the main purpose of

conceptualizing a speciﬁc object of study. The term

Ontology has its origin in philosophy and denotes the

philosophical discipline that deals with the nature and

Pinto, V. and Parreiras, F.

OntoDIVE: An Ontology for Representing Data Science Initiatives upon Big Data Technologies.

DOI: 10.5220/0009416500420051

In Proceedings of the 22nd International Conference on Enterprise Information Systems (ICEIS 2020) - Volume 1, pages 42-51

ISBN: 978-989-758-423-7

the organization of reality (Osterwalder, 2004). In this

sense, ontology involves identifying the fundamental

categories of things (Parreiras, 2011).

Although there exist many deﬁnitions of ontolo-

gies in the scientiﬁc literature, some elements are

common to these deﬁnitions: a computer ontology

is said to be an agreement about a shared, formal,

explicit and partial account of a conceptualisation

(Spyns et al., 2002). The common vocabulary of an

ontology, deﬁning the meaning of terms and their re-

lations, is usually organized in a taxonomy and con-

tains modeling primitives such as concepts, relations,

and axioms (Staab and M

adche, 2000). Thus, ontolo-

gies constitute formal models of some aspect of the

world that may be used for drawing interesting log-

ical conclusions even for large models (Staab et al.,

2010). For (Spyns et al., 2002), an ontology contains

the vocabulary (terms or labels) and the deﬁnition of

the concepts and their relationships for a given do-

main. In many cases, the instances of the application

(domain) are included in the ontology as well as do-

main rules (e.g. identity, mandatoriness, rigidity, etc.)

that are implied by the intended meanings of the con-

cepts. The role of ontologies is to capture domain

knowledge and provide a commonly agreed upon un-

derstanding of a domain (Staab and M

adche, 2000).

For (Mizoguchi and Ikeda, 1998), the purpose of on-

tology engineering is to provide a basis of building

models of all things in which computer science is in-

terested.

This study intends to answer the following re-

search question: ”How can we represent a initiative

of data science upon big data?”. By addressing this

research question, this study presents OntoDIVE - an

ontology to explain interactions between people, pro-

cesses and technologies in a data science initiative

upon big data. This study is structured as follows:

Section 2 presents methods of this research. In sec-

tion 3 research results are presented. These results

are discussed in section 4. This study is concluded

and future work is suggested in section 5.

2 METHODS

In order to address research question, we followed

Knowledge Meta Processes guidelines from Ontol-

ogy Engineering Methodology, which consists of ﬁve

main steps: a) Feasibility Study, b) Kickoff, c) Re-

ﬁnement, d) Evaluation and e) Application and Evo-

lution (Sure et al., 2009). For the ﬁrst step, a group of

seven professionals involved with data science upon

big data initiatives in a large organization was inter-

viewed in order to capture their thoughts. This group

was asked the following questions: a) To what extent

would the lack of vocabulary prevent data science ini-

tiatives over big data? b) To what extent could an ar-

tifact as an ontology leverage data science initiatives

over big data? Insights were compiled and a technical

speciﬁcation was generated. Then, software Prot

Desktop version 5.5.0 was used to build an ontology

as a representational instrument (Musen, 2015). DL

Query plugin was installed on top of Prot

e to vali-

date ﬁnal ontology. This plugin is based on Manch-

ester OWL syntax and presents a user-friendly syntax

for OWL DL.

3 RESULTS

In this section we present results for each stage of this

research. Subsection 3.1 presents collected insights

for problems, opportunities and potential solutions.

A summary of ontology requirements is presented in

subsection 3.2, including competency questions sup-

posed to be addressed by ﬁnal ontology. Subsection

3.3 presents general information about the ontology

designed to meet identiﬁed requirements. Subsection

3.4 presents answers to competency questions raised

during requirements gathering stage.

3.1 Feasibility Study

Professionals selected for this study were asked to

what extent the lack of vocabulary could prevent data

science initiatives over big data. According to them,

one of the major problems related to data science

upon big data is that companies are still struggling to

choose big data technologies and to deﬁne which data

science framework to choose. Such decisions involve

multiple concepts not always dominated by internal

personnel and part of consultants have exclusive part-

nerships with technology developers or standard pub-

lishers. For them, creating an ontology capable of in-

tegrating People, Processes and Technologies would

help to establish a common ground for all parts in-

volved in the whole big data ecosystem.

Respondents were also asked to what extent an on-

tology could leverage data science initiatives upon big

data. According to them, there is an opportunity for

any kind of artifact capable of clariﬁcation of con-

cepts. They believe business challenges can be ad-

dressed by data science applications deployed as ﬁnal

result of data science processes carried out over big

data technologies. Nevertheless, they believe such

initiatives are performed by people and if everyone

involved clearly understands the concepts and termi-

OntoDIVE: An Ontology for Representing Data Science Initiatives upon Big Data Technologies

People

ProcessTechnology

Ontology for Data

Science Initiatives

upon Big Data

Source: Authors.

Figure 1: Semi-formal Description of the Ontology.

nologies then the dissemination and usage of data sci-

ence upon big data will be facilitated.

3.2 Kickoff

In a general way, the expected ontology should act as

an information provider on the existing relationship

between people, processes and technologies in data

science initiatives upon big data. A semi-formal de-

scription of expected ontology is represented in Fig-

ure 1. Next subsections present summary of ontology

requirements speciﬁcation document.

3.2.1 Ontology Purpose

Purpose of ontology will be to provide a knowledge

model of data science initiatives upon big data tech-

nologies.

3.2.2 Ontology Scope

Ontology will focus on relationships between people,

processes and technologies during data science initia-

tives upon big data technologies.

3.2.3 Implementation Language

Ontology will be implemented in Proteg

e as it is

based on Java and supports the latest OWL 2 Web

Ontology Language and RDF speciﬁcations from the

World Wide Web Consortium (W3C). DL Query plu-

gin will be used for ontology validation.

3.2.4 Intended End Users

Ontology will be directed to end users below.

• Customers: organizations from any economic

activity who want to run data science initiatives

upon big data technologies in their own opera-

tions.

• Consultants: people or group of people who

want to provide consultancy services for data sci-

ence initiatives upon big data technologies

• Technology Developers: people or group of peo-

ple who develop big data technologies and want to

analyze how their products ﬁt in real-world data

science initiatives

• Standards Publishers: people or group of people

who develop data science processes and want to

analyze usage of their processes

• Researchers: people or group of people who

want to perform comprehensive analyses of data

science initiatives upon big data technologies for

academic purposes

3.2.5 Intended Uses

Ontology will be designed to use cases below.

• UC01: Customers need to search for prior data

science upon big data and also for related people,

processes and technologies

• UC02: Consultants need to publish prior experi-

ences related to data science initiatives upon big

data

• UC03: Technology Developers need to publish

new technologies, functionalities and upgrades

• UC04: Standards Publishers need to publish new

standards, approaches and/or methods

• UC05: Researchers need to provide statistics and

explain interactions of people, processes and tech-

nologies

3.2.6 Non Functional Requirements

Ontology will address non functional requirements

below.

• NFR01: The ontology must support a multilin-

gual scenario

• NFR02: The ontology must be based on de facto

standards in existence.

3.2.7 Functional Requirements

Ontology will address particular functional require-

ments for Customers (CQ01, CQ02 and CQ03), Con-

sultants (CQ04, CQ05 and CQ06), Technology De-

velopers (CQ07), Standards Publishers (CQ08) and

Researchers (CQ09 and CQ10).

• CQ01: What are the internal functional processes

ready for a data science initiative upon big data?

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

Table 1: OntoDIVE - Object Properties Related to People.

Property Domain Range Usage

afﬁliationOf Afﬁliations People, Processes and

Technologies

This property relates afﬁliation (such as.:

companies, communities, associations) to

other classes.

havePriorExperience People Processes, Tech-

nologies, Functions,

Positions and Softskills

This property relates people to their expe-

riences

haveSoftSkills People SoftSkills This property relates people to soft skills

haveFunction People Function This property relates people to functions

havePosition People Positions This property relates positions to people

locatedAt People Locations This property relates locations to people

peopleLinking People People This property relates an instance from class

people to another instance of the same

class

Source: Authors.

Table 2: OntoDIVE - Object Properties Related to Processes.

Property Domain Range Usage

processOf Processes Frameworks This property relates processes to frameworks

activityOf Activities Processes This property relates activities to other classes

artifactOf Artifacts Processes This property relates artifacts to other classes

constraintOf Constraints Processes This property relates constraints to other classes.

Processes may have constraints

performanceOf Performance Processes This property relates performance metrics to other

classes

changeEventOf ChangeEvent Frameworks

Technologies

This property change events to frameworks and

technologies

targetOf Targets Performance This property relates targets to other classes

lifeCycleOf LifeCycle Frameworks

Technologies

This property relates lifecycle to other frameworks

and technologies

frameworkLinking Processes Processes This class relates a particular class to itself

Source: Authors.

• CQ02: What are the technologies internally avail-

able for starting a data science initiative upon big

data?

• CQ03: What are the internal people (or group of

people) skilled for a data science initiative upon

big data?

• CQ04: Which data science frameworks are rec-

ommended by a particular consultant?

• CQ05: Which big data technologies are recom-

mended by a particular consultant?

• CQ06: What are the people skilled for a data sci-

ence initiative upon big data working for a partic-

ular consultant?

• CQ07: Which algorithms, methods or techniques

are included into a particular technology?

• CQ08: Which activities are expected to be per-

formed during a data science initiative according

to a particular framework?

• CQ09: What are known data science frameworks

available in the market?

• CQ10: What are known big data technologies

available in the market?

3.3 Reﬁnement

The ontology built in this study was named as On-

toDIVE, a short for Ontology for data science Ini-

tiatives upon big data technologies. The core class

of OntoDIVE is DSUponBD which is intended to be

applicable to a very broad range of big data initia-

tives. This class represents a data science initiative

upon big data. The main purpose of this class is to

represent integration between people, processes and

technologies. Each instance of this class represents a

single initiative and each initiative may be related to

OntoDIVE: An Ontology for Representing Data Science Initiatives upon Big Data Technologies

Table 3: OntoDIVE - Object Properties Related to Technologies.

Property Domain Range Usage

algorithmOf Algorithms Techniques, Methods

and Technologies

This property relates algorithms to other

classes

methodOf Methods Technologies, Ap-

proaches, Techniques

and Algorithms

This property relates methods to other

classes

techniqueOf Techniques Methods, Algorithms

and Technologies

This property relates techniques to other

classes

approachOf Approaches Methods, Technologies,

Processes and Theories

This property relates approaches to other

classes

theoryOf Theories Approaches and Tech-

nologies

This property relates academic theories

to classes

categoryOf Categories Roles This property relates categories to roles,

according to taxonomy of big data

haveRole Technologies Roles This property relates technologies to

roles, according to taxonomy of big data

technologies

haveCategory Roles Categories This property relates technologies to cat-

egories, according to big data taxonomy

technologiesLinking Technologies Technologies This property relates a particular class to

itself

Source: Authors.

other initiatives. Class People is related to soft skills,

positions, functions, locations, etc. People may be

related to other people. Class Frameworks repre-

sents all processes used to support data science ini-

tiatives. This class is related to: processes, activi-

ties, resources, technologies, people, afﬁliations, the-

ories, etc. Frameworks may be related to other pro-

cesses. Class Technologies describes all technolo-

gies involved in a data science initiative upon big

data. This class is related to: roles, theories, ap-

proaches, methods, techniques, algorithms, etc. Fig-

ure 2 shows all classes OntoDIVE intends to explain

and their relationships. Next subsections present de-

tails of classes and properties, grouped by people,

processes and technologies.

3.3.1 People Perspective

Classes below are related to people.

• People: This class describes all people involved

in a data science Initiative. A particular initia-

tive may require relationship between different

people. A data engineer may be associated to a

project leader, for example.

• Afﬁliations: This class explains afﬁliations such

as: organization, communities, associations, com-

panies, etc. Should a particular data science ini-

tiative requires accredited personnel from speciﬁc

organizations, this class could be used.

• Functions: This class describes functions per-

formed by people in data science initiatives. Peo-

ple may perform different functions in different

initiatives. One may be data scientist in a particu-

lar initiative and a data engineer in another initia-

tive.

• Positions: This class describes positions occu-

pied by people in data science Initiatives. A par-

ticular initiative may require a hierarchical or a

matrix structure. This class supports this kind of

relationship. Some examples include: director,

manager, staff, among others.

• SoftSkills: This class describes soft skills re-

lated to data science initiatives. An initiative may

require certain skills, such as: logical thinking,

communication, etc.

• Locations: This class describes physical loca-

tions where people are based at. People may be

physically based in a different location of their or-

ganizations.

3.3.2 Processes Perspective

Classes below are related to processes.

• Frameworks: This class describes frameworks

to support data science initiatives upon big data.

Each initiative may adopt a different framework.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

Source: Authors.

Figure 2: OntoDIVE Overview.

Examples of frameworks include: CRISP-DM,

KDD, etc

• LifeCycle: This class represents a stage of a par-

ticular framework or technology. An initiative

may use a framework during its experimental life-

cycle. CRISP-DM, as an example, may evolve to

CRISP-DM 2.0

• ChangeEvent: This class represents an event

which resulted in a change of a particular frame-

work. New version of a set of best practices is an

example of this kind of event.

• Processes: This class describes processes associ-

ated to frameworks. Each framework has partic-

ular processes, such as: business understanding,

data selection, etc.

• Activities: This class describes activities related

to a particular process. Process ”data selection”

includes some activities such as: identify data-

sources, acquire data, etc.

• Artifacts: This class describes artifacts required

or generated by a process. All processes generate

outputs based on their inputs. Artifacts may be

input or output of processes.

• Constraints: This class describes constraints that

restrict processes. Some examples include: lan-

guage to use, computational environment, etc.

• Resources: This class describes resources re-

quired by processes. People and technologies are

some examples.

• Performance: This class is used to clarify the

goals of end user in terms of what he wants to

obtain from data.

3.3.3 Technologies Perspective

Classes below are related to technologies.

• Technologies: This class describes technologies

of data science Initiatives.

• Roles: This class explains roles performed by

technologies according to taxonomy of big data

technologies. Few examples: data creation, data

acquisition, etc.

• Categories: This class explains categories of

technology roles. Role ”data creation” may be

categorized into sensors, logs, etc.

• Theories: This class describes all theories related

to a particular technology or framework. Some

examples include: information theory, automata

theory, database theory, machine learning theory,

etc

• Approaches: This class describes approaches re-

lated to theories, frameworks or technologies. A

particular initiative may require an speciﬁc ap-

proach which may bring a speciﬁc set of tech-

niques, methods and algorithms. Machine learn-

ing theory has some approaches: supervised

learning, unsupervised learning, reinforcement

learning, multi-task learning, etc.

• Techniques: This class describes techniques re-

lated to a technology or approach. Supervised

learning approach may be implemented by clas-

siﬁcation or by regression.

• Methods: This class describes methods related to

a technology or process. Supervised learning by

classiﬁcation may be implemented by rule learn-

OntoDIVE: An Ontology for Representing Data Science Initiatives upon Big Data Technologies

ing, neural networks, support vector machines,

etc.

• Algorithms: This class describes algorithms.

Neural networks for classiﬁcation may be imple-

mented as Radial Basis Function, Incremental Ra-

dial Basis Function, etc.

3.4 Evaluation and Application

This section describes both the queries designed to

answer competency questions presented in section 3.2

and the outputs provided by OntoDIVE after execu-

tion of those queries. It is important to highlight that

classes and properties of OntoDIVE were populated

in small scale as the purpose of this study is to build

and validate an ontology capable of explain data sci-

ence initiatives upon big data technologies. Although

OntoDIVE has been designed to explain interactions

it has not been applied in productive systems.

For each functional requirement written in natu-

ral language and described in section 3.2.7, there is a

corresponding ﬁgure presenting DL Query and Onto-

DIVE outputs. Figures 3, 4 and 5 address competency

questions from customers. Figures 6, 7 and 8 address

competency questions from consultants. Figure 9 ad-

dresses competency questions from technology de-

velopers. Figure 10 addresses competency questions

from standard publishers. Figures 11, 12 address

competency questions from researchers.

Source: Authors.

Figure 3: OntoDIVE outputs for CQ01.

Source: Authors.

Figure 4: OntoDIVE outputs for CQ02.

Source: Authors.

Figure 5: OntoDIVE outputs for CQ03.

4 DISCUSSION

OntoDIVE may be used to explain interactions be-

tween people, processes and technologies in a data

science initiative upon big data. The more populated

OntoDIVE is the more accurate it will be to answer

relevant competency questions. These answers could

help organizations from all segments to leverage their

data science initiatives over big data. Thus, Onto-

DIVE could be used as basis of a framework to sup-

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

Source: Authors.

Figure 6: OntoDIVE outputs for CQ04.

Source: Authors.

Figure 7: OntoDIVE outputs for CQ05.

port organizations in their initiatives.

OntoDIVE may also be used to conﬁrm interdisci-

plinary nature of data science initiatives upon big data

technologies. Considering only instances created for

this study, it is possible to see ﬁve different proﬁles of

data scientist: domain expert, statistician expert, com-

puting expert, business expert and communicator ex-

pert. Additionally, there are other relevant functions

such as data engineer and project manager. In this

study, each function was performed by different per-

sonnel with different academic background.

Source: Authors.

Figure 8: OntoDIVE outputs for CQ06.

Source: Authors.

Figure 9: OntoDIVE outputs for CQ07.

Another potential usage for OntoDIVE is to make

organizations capable of drawing parallels between

data science results for a particular domain to their

own domain. As an example, mining industry pro-

fessionals could use OntoDIVE to clearly understand

how data science initiatives are conducted in health-

care or transportation industries. This has a poten-

tial to enlarge possibilities of applications. Onto-

DIVE may be used as a tool for comparison of both

data science frameworks and big data technologies.

As each framework has strengths and weaknesses, it

OntoDIVE: An Ontology for Representing Data Science Initiatives upon Big Data Technologies

Source: Authors.

Figure 10: OntoDIVE outputs for CQ08.

Source: Authors.

Figure 11: OntoDIVE outputs for CQ09.

is important for customers to choose the more suit-

able for them taking into account internal restrictions.

Whithin the same rationale, OntoDIVE may be used

as a tool for comparison of big data technologies. On-

toDIVE brings light to the fact that sometimes cus-

tomers are presented to a reduced list of frameworks

or technologies only because a particular consultant

does not work with such framework or technology. In

this regard, OntoDIVE may be used to protect cus-

tomers of being subject to commercial interests de-

pending on consultants they rely on.

Source: Authors.

Figure 12: OntoDIVE outputs for CQ10.

Main goal of mining industry is to minimize the

amount of assets and resources required to run opera-

tions. There are many opportunities but, in a general

way, industry seems to be focusing on cost drivers,

such as: 1) increase productivity; 2) increase prof-

itability; 3) improve assets management. In this con-

text, data science initiatives upon big data should be

focused on understanding how to reduce waste in sup-

ply chain and on ﬁnding what precisely drives fuel

consumption. In the speciﬁc case of mining company

object of this study, Figure 3 shows internal processes

with data science readiness. This readiness is related

to existence of systems that capture and make process

data available. Figure 4 shows technologies internally

available for starting a data science initiative. Min-

ing industries usually have condition monitoring sys-

tems, dispatch and fatigue management systems for

mining operations. A data science initiative should

consider gathering data from these systems. Figure

5 shows people with knowledge or prior experience

in data science initiatives. Although people may be

geographically dispersed in a large company, it is rel-

evant to identify everyone that could add value to a

data science initiative.

5 CONCLUSION

This work contributes to the clariﬁcation of concepts

and terminologies related to data science and big data.

Data science initiatives upon big data can be analyzed

from three different perspectives: people, processes

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

and technologies. OntoDIVE ontology, proposed in

this study, explains relationships among these terms

and concepts in the context of a data science initiative

upon big data. Proposed ontology may also be used

to conﬁrm interdisciplinary nature of data science ini-

tiatives upon big data technologies. Another potential

usage for OntoDIVE is to make organizations capable

of drawing parallels between data science results for a

particular domain to their own domain. The ontology

may be used as a tool for comparison of both data sci-

ence frameworks and big data technologies and could

be used as basis of a framework to support organiza-

tions in their data science initiatives upon big data.

This work has several limitations. Although On-

toDIVE was designed to be a comprehensive arti-

fact capable of explaining most part of data science

initiatives upon big data, it was built based on con-

siderations from specialized professionals that work

for a single organization. Furthermore, while sev-

eral ontologies within the same domain are developed

independently by different communities, this study

was never focused on merging OntoDIVE with exist-

ing ontologies and no method was used for ontology

alignment, as proposed by (Idoudi et al., 2016). Ad-

ditionally, while OntoDIVE was validated by descrip-

tion logics queries it has not been applied in produc-

tive system. Therefore, OntoDIVE should be consid-

ered as an initial version.

Future works could create a friendly graphical

user interface to allow interaction with OntoDIVE,

since Proteg

e interface is not understood by many

who are not knowledgeable about ontologies and their

editing tools. Future works coud also apply Onto-

DIVE on productive systems in order to collect in-

sights and thoughts of more people and organizations

focusing on evolution of the ontology.

REFERENCES

Abdullah, M. F., Ibrahim, M., and Zulkiﬂi, H. (2017). Re-

solving the misconceptions on big data analytics im-

plementation through government research institute in

malaysia. In International Conference on Internet of

Things, Big Data and Security, volume 2, pages 261–

266. SCITEPRESS.

Dietrich, D., Heller, B., and Yang, B. (2015). Data Science

and Big Data Analytics: Discovering, Analyzing, Vi-

sualizing and Presenting. Wiley, [s.l].

Hurwitz, J., Kaufman, M., and Bowles, A. (2015). Cogni-

tive Computing and Big Data Analytics. Wiley Pub-

lishing, [s.l], 1st edition.

Idoudi, R., Ettabaa, K. S., Hamrouni, K., and Solaiman, B.

(2016). Fuzzy clustering based approach for ontology

alignment. In ICEIS (1), pages 594–599.

Liu, J., Li, J., Li, W., and Wu, J. (2016). Rethinking big

data: A review on the data quality and usage issues.

ISPRS Journal of Photogrammetry and Remote Sens-

ing, 115:134 – 142. Theme issue ’State-of-the-art in

photogrammetry, remote sensing and spatial informa-

tion science’.

Mizoguchi, R. and Ikeda, M. (1998). Towards ontology

engineering. Journal-Japanese Society for Artiﬁcial

Intelligence, 13:9–10.

Musen, M. A. (2015). The prot

E project: A look back

and a look forward. AI Matters, 1(4):4–12.

Osterwalder, A. (2004). The business model ontology a

proposition in a design science approach. PhD the-

sis, Universit

e de Lausanne, Facult

e des hautes

etudes

commerciales.

Parreiras, F. S. (2011). Marrying model-driven engineering

and ontology technologies: the twouse approach.

Spyns, P., Meersman, R., and Jarrar, M. (2002). Data mod-

elling versus ontology engineering. SIGMOD Rec.,

31(4):12–17.

Staab, S. and M

adche, A. (2000). Axioms are Objects, too

- Ontology Engineering beyond the Modeling of Con-

cepts and Relations. In Benjamins, V., Gomez-Perez,

A., and Guarino, N., editors, Proceedings of the Work-

shop on Applications of Ontologies and Problem-

solving Methods, 14th European Conference on Ar-

tiﬁcial Intelligence ECAI 2000, Berlin, Germany.

Staab, S., Walter, T., Gr

oner, G., and Parreiras, F. S. (2010).

Model Driven Engineering with Ontology Technolo-

gies, pages 62–98. Springer Berlin Heidelberg, Berlin,

Heidelberg.

Sure, Y., Staab, S., and Studer, R. (2009). Ontology en-

gineering methodology. In Handbook on ontologies,

pages 135–152. Springer.

Vera-Baquero, A., Colomo-Palacios, R., and Molloy, O.

(2016). Real-time business activity monitoring and

analysis of process performance on big-data domains.

Telematics and Informatics, 33(3):793 – 807.

OntoDIVE: An Ontology for Representing Data Science Initiatives upon Big Data Technologies