IOSR Journal of Mobile Computing & Application (IOSR-JMCA)
e-ISSN: 2394-0050, P-ISSN: 2394-0042.Volume 3, Issue 4 (Jul. - Aug. 2016), PP 52-61
www.iosrjournals.org
DOI: 10.9790/0050-03045261 www.iosrjournals.org 52 | Page
Data Scientist Skills
Nur Amie Ismail, Wardah Zainal Abidin
(Advance Informatics School, University of Technology Malaysia, Kuala Lumpur)
Abstract: Decision making is one of the most important aspects in order to enhance service delivery to citizens
and businesses, gain more profit, and help stakeholders to strategize their business functions. Nowadays, most of
the stakeholders make decisions based on the data that is precise, concise, appropriate, and accurate. Even though
Big Data Analytic (BDA) tools and software can assist in this matter, skills and competency of the personnel that
handle and manage the data is more crucial and important. Thus, the aim of this paper is to identify data scientist
skills from global best practices and examine the most important data scientist skills required by Information
Technology (IT) personnel. From our findings, we found 44 data scientist skills and the top 5 (five) skills are
business, statistic, machine learning, communication, and analysis.
Keywords: Data Science, Data Scientist, Data Scientist Skill
I. Introduction
Nowadays with the vast amounts of data available in the world, companies across industry are focusing
on exploiting data for their competitive advantage. Hence, they realized that they need to hire more data scientists
or equip their employees with data scientist skills. Data scientist is an expert who is capable to extract meaningful
value from the data and also manage the whole lifecycle of data [1]. Data scientists also help to bridge the
communication gap between business and IT functions, proposing meaningful measures, modelling the data,
visualizing the output, sharing the technique, and automating the process [2]. According to McKinsey Global
Institute, in United Stated of America (USA) alone, they need another 140 - 190 thousand of data scientist by
2018. Whereas in Malaysia, Multimedia Development Corperation (MDeC) have a set an ambitious target to
produce 1,500 data scientists by 2020. According to [3], currently in Malaysia there are only eighty (80) data
scientist across the country. In order to increase number of data scientist, various programs have been arranged
such as Big Data conference, trainings, certification, and Massive Open Online Courses (MOOC). However,
programs that have been arranged still insufficient to cater for meeting the ambitious target. Thus, this paper will
identify data scientist skills from global best practices and examine the most important data scientist skills required
by IT personnel in order to be recognized as Data Scientist.
This paper is organized as per following sequence. Section 2 explained review methodology that has
been used in this study. Section 3 briefly explained data science definition, data science fundamental concept, and
the difference between Data Scientist and Data Analyst. Section 4 discusses the data scientist skills from global
best practices followed by section 5 about finding. Finally, the conclusion and future works is in Section 6.
II. Review Methodology
To get clear and better understanding on the research topic, research review has been conducted from
various resources such as books, articles, journals, and web sites. List of computerized databases used in this paper
are Association for Computing Machinery (ACM) Digital Library, IEEE, Science Direct, Springer Link, Wiley
Online Library, ERIC, Gartner, and Google Scholar. The resources that are reviewed are within the period of 2011
to 2016. Fig. 1 below shows the review methodology that has been used in this paper:
Fig. 1: Research review methodology
Data Scientist Skills
DOI: 10.9790/0050-03045261 www.iosrjournals.org 53 | Page
III. Data Science Definition, Data Science Fundamental Concepts And Difference Between
Data Scientist And Data Analyst
3.1 Data Science Definition
There are several definitions of data scientists from several authors as listed in the table 1.
Table 1: Data Science Definitions
No.
Definition
1
Set of fundamental principles that support and guide the principle extraction of information and
knowledge from data [4].
2
Data science is the study of the generalizable extraction of knowledge from data [5].
3
Data science is a combination of statistic, computer science, and information design [2].
From table 1, we can summarize that, data science is combination of field of study related to extraction and
transformation of data.
3.2 Data Science Fundamental Concepts
According to [2], the fundamental concept of data science is extracting useful knowledge from data to
solve business problems that can be treated systematically by following a process with reasonably well-defined
stages. Data-science results requires careful consideration of the context in which they will be used in the
relationship between the business problem and the analytics solution. This often can be decomposed into tractable
sub problems via the framework of analyzing expected value. IT can be used to find informative data items from
within a large body of data. Other than that, entities that are similar with respect to known features or attributes
often are similar with respect to unknown features or attributes, data might not generalize beyond the observed
data and to draw causal conclusions, an attention to the presence of confounding factors possibly unseen ones.
3.3 Difference Between Data Scientist And Data Analyst
According to [6], Data Analyst focus on the movement and interpretation of data, typically focus on the
past and present. Where Data Scientist focus on summarizing data and to provide forecasting based on pattern
identified from past and current data. [7] define and differentiate between Data Scientist and Data Analyst as
describe in table 2.
Table 2: Data Scientist vs Data Analyst
Data Scientist
Data Analyst
Building statistical models that make decisions
based on data. Each decision can be hard, e.g. block
a page from rendering, or soft, e.g. assign a score for
the maliciousness of a page that is used by
downward systems or humans.
Writing custom queries to answer
complex business questions.
Conducting causality experiments that attempt to
attribute the root cause of an observed phenomenon.
This can be done by designing A/B experiments or
if A/B experiment is not possible apply
epidemiological approach to the problem.
Conceiving and implementing new
metrics on capturing previously
poorly understood parts of the
business / product.
Identifying new products or features that come from
unlocking the value of data; being a thought leader
on the value of data. A good example of that is the
product recommendations feature that Amazon first
made available to a mass audience.
Addressing data quality issues, such
as data gaps or biases in data
acquisition. Working with the rest
of engineering to instrument
incremental new data acquisition.
IV. Data Scientist Skills
In order to explore the list of data scientist skills, this study has a global reach and perspective as well
includes the Malaysian public sector. Basically the data scientist incorporates advanced analytical approach using
sophisticated analytic and data visualization software or tools in order to discover patterns of the data. The data
scientist then will do data migration and integration, data cleaning, analyzing and deliver the outcomes. According
to [8], the data scientist must be able to write in different programming language such as Python, R, Java, Ruby,
Clojure, Matlab, Pig, and SQL. Other than that, the data scientist need to understand about Hive, Hadoop, and
Map Reduce. They also suggest that the data scientist must be familiar with Natural Language Processing (NLP),
machine learning, conceptual modelling, statistical analysis, predictive modelling, and hypothesis testing. Even
Data Scientist Skills
DOI: 10.9790/0050-03045261 www.iosrjournals.org 54 | Page
though the data scientist has to learn new skills as explained above, at least they should have the capabilities in
communication skills, querying the database, understand about business strategy, able to design simple prototype
for top management, and have good understanding in system architecture.
Educational data scientist is rarely sighted breed especially within business and government. In order to
tackle this scenario, we need to produce more graduates and also equip the employees with necessary skills in
data science. [2] suggest that the data scientist should have skills in data mining, data modelling, data visualization,
and machine learning. According to [9], the data scientist uses advanced analytics such as predictive analysis, data
visualization and modelling, and machine learning to predict what is going to happen in the future and give
recommendations to enhance existing business process. They also defines that the data scientist is a combination
of three(3) main fields which are computer science, statistics, and domain knowledge.
Fig. 2 shows the relationship and skills for each area.
Fig. 2:Data scientist skills (Ayankoya et al., 2014)
[5] emphasizes that machine learning is the most important skill and necessary for all data scientists. In
machine learning, the data scientist should master of all 3 class of skills as illustrated in table 3 as per below.
Other than machine learning, data scientist also required knowledge in text mining, markup language like XML,
mathematics, and artificial intelligence (AI).
Table 3: Three (3) class of skill in Machine Learning
No.
Class
Skills
1
Statistic
Bayesian statistic and probability.
2
Computer Science
Data structure, algorithm and distributed computer.
3
Correlation and Causation
Modelling.
Data scientist is an expert that has the ability to manipulate and extract knowledge and turn it into
meaningful value [1]. According to their study, currently there is no accepted and effective data science
professional curriculum. [10] found that the two (2) top skills companies are looking for in a data scientist are
programming and statistical. The details of these two (2) skills illustrate in fig. 3.
Fig. 3: Programming and Statistics
According to [11], data science is now being benchmarked against practices that employed on highly
skilled professionals. Data Scientist uses scientific methods to discover knowledge and patterns of the data.
Data Scientist Skills
DOI: 10.9790/0050-03045261 www.iosrjournals.org 55 | Page
Fig. 4: The data science benefits-realization process(Viaene, 2013)
Fig. 4 illustrates on how to use data to improve the business. This process involves modelling,
discovering, operationalizing, and cultivating the knowledge. The data scientist must have a pretty good skill in
business domain, analysis, and communication. While, according to [12], data scientist is the sexist job in this
century. Sexy in the sense of having a rare quality in high demand. Data scientist is urgently needed by
organizations because they know how to use the analysis of big data to make effective decisions. Among the skill that
they should consider are programming language, computer science, mathematics, economics, probability, and business. In
the O’Reilly book, Analysing the Analysers by [13], they have made a survey over more than 200 data scientists
to discover and analyze what data skills needed by the data scientist. They found 22 generic skills shown in fig.
5.
Fig. 5: Data Scientist generic skills
Malaysia Public Sector have started their BDA project since 2014 led by MAMPU. Since data science is
still new in Malaysia, they do not have internal expertise in this area. Therefore, they have hired external
consultants to develop the project. Even though they use external parties; knowledge transfer, training, and
Data Scientist Skills
DOI: 10.9790/0050-03045261 www.iosrjournals.org 56 | Page
technology updates are given to Government IT officers. Soon, Malaysian Government do realize that the
importance of having internal expertise in this field. Hence, skills of data scientist are identified to enhance
Government IT officer competency and knowledge. According to [14], the skills that are required for data scientist
consist of model and analysis, data processing, statistic, business domain, soft skill, and technical skill as
illustrated in Fig. 6.
Fig. 6: Data scientist skill (Suhailis, 2016)
In 2013, [15] have announced the Digital Malaysia Roadmap, which encompasses a plan that addresses
three ICT areas which are to access, adoption and usage ICT services. One of the goals in the roadmap is to
improve Big Data literacies in Malaysia. Therefore, in October 2013, MDeC have conducted a survey to 17
experts in Big Data. The participants come from different background such as telecommunication company,
universities, marketing agency, software development companies, and others. Based on their survey, the top five
skills needed are:
(i) Big and Distributed Data (eg: Hadoop, MapReduce)
(ii) Algorithms (eg: computational complexity, CS theory)
(iii) Machine Learning (eg: decision trees, neural nets, SVM, clustering)
(iv) Back-End Programming (eg: JAVA/Rails/Objective C)
(v) Visualization (eg: statistical graphics, mapping, web-based dataviz)
In the last few years, the interest in data science field has soared. Most of the companies in USA are
seeking and recruiting employees who have skills related to data science. From the perspective of [16], she
emphasizes that the data scientist must have both technical skill and non-technical as listed in the table 4 below:
Table 4: List of skill needed to recruit employee in data science
No.
Type of skill
Skills
1
Technical Skills
Analytics, SAS, R, Python, Coding, Hadoop, SQL, and Database.
2
Non-Technical Skills
Intellectual curiosity, business acumen, and communication skills.
In United Kingdom, data science is among the most rapidly emerging field based on trend in ICT market.
The key to success in business nowadays is to understand customer’s preferences, needs, and behavior. Thus, data
scientist plays an important role to do a prediction and make decision in this particular area. [17] concludes that
data scientist need multi-faceted skills illustrated in fig. 7.