Big data sources and methods for social and economic analyses

Contents lists available at ScienceDirect
Technological Forecasting & Social Change
journal homepage: www.elsevier.com/locate/techfore
Big Data sources and methods for social and economic analyses
Desamparados Blazquez, Josep Domenech*
Department of Economics and Social Sciences, Universitat Politècnica de València, Camí de Vera s/n., Valencia 46022, Spain
A R T I C L E I N F O
Keywords:
Big Data architecture
Forecasting
Nowcasting
Data lifecycle
Socio-economic data
Non-traditional data sources
Non-traditional analysis methods
A B S T R A C T
The Data Big Bang that the development of the ICTs has raised is providing us with a stream of fresh and
digitized data related to how people, companies and other organizations interact. To turn these data into
knowledge about the underlying behavior of the social and economic agents, organizations and researchers must
deal with such amount of unstructured and heterogeneous data. Succeeding in this task requires to carefully plan
and organize the whole process of data analysis taking into account the particularities of the social and economic
analyses, which include the wide variety of heterogeneous sources of information and a strict governance policy.
Grounded on the data lifecycle approach, this paper develops a Big Data architecture that properly integrates
most of the non-traditional information sources and data analysis methods in order to provide a specifically
designed system for forecasting social and economic behaviors, trends and changes.
1. Introduction
What comes to your mind when talking about “The Digital Era”? For
sure, concepts as the “Internet”, “Smartphones” or “Smart sensors”
arise. These technologies are progressively being used in most of the
everyday activities of companies and individuals. For instance, many
companies conduct marketing campaigns through social networks, sell
their products online, monitor the routes followed by sales re-
presentatives with smartphones or register the performance of ma-
chinery with specific sensors. At the other side, individuals make use of
computers, smartphones and tablets in order to buy products online,
share their opinions, chat with friends or check the way to some place.
Moreover, citizens' movements and activities are daily registered by
sensors placed in any part of cities or roads and in public places such as
supermarkets.
Therefore, all of these technologies are generating tons of digitized
and fresh data about people and firms' activities that properly analyzed,
could help reveal trends and monitor economic, industrial and social
behaviors or magnitudes. These data are not only updated, but also
massive, given that daily data generation has been recently estimated in
2.5 Exabytes (IBM, 2016). For this reason, they are commonly referred
to as “Big Data”, concept which first appeared in the late 90s (Cox and
Ellsworth, 1997) and was defined in the early 2000s in terms of the 3Vs
model (Laney, 2001), which refers to: Volume (size of data), Velocity
(speed of data transfers), and Variety (different types of data, ranging
from video to data logs for instance, and with different structures). This
model evolved to adapt to the changing digital reality, so that it was
extended to 4Vs, adding the “Value” dimension (process to extract
valuable information from data, known as Big Data Analytics). Cur-
rently, the “Big Data” concept is starting to be defined in terms of the
5Vs model (Bello-Orgaz et al., 2016), which added the “Veracity” di-
mension (related to proper data governance and privacy concerns).
This new data paradigm is called to transform the landscape for
socio-economic policy and research (Einav and Levin, 2014; Varian,
2014) as well as for business management and decision-making. Thus,
identifying which data sources are available, what type of data they
provide, and how to treat these data is basic to generate as much value
as possible for the company or organization. In this context, a Big Data
architecture adapted to the specific domain and purpose of the orga-
nization contributes to systematize the process of generating value. This
architecture should be capable of managing the complete data lifecycle
in the organization, including data ingestion, analysis and storage,
among others.
Furthermore, the design of a Big Data architecture should consider
the numerous challenges that this paradigm implies. These include:
scalability, data availability, data integrity, data transformation, data
quality, data provenance (related to generation of right metadata that
identify the origin of data as well as the processes applied to them
during the data lifecycle, to assure traceability), management of huge
volumes of information, data heterogeneity (structured and un-
structured, with different time frequencies), integration of data from
different sources, data matching, bias, availability of tools for properly
analyzing such kind of data, processing complexity, privacy and legal
issues, and data governance (Fan et al., 2014; Jagadish et al., 2014;
http://dx.doi.org/10.1016/j.techfore.2017.07.027
Received 1 March 2017; Received in revised form 7 July 2017; Accepted 25 July 2017
*
Corresponding author.
E-mail addresses: mdeblzso@upvnet.upv.es (D. Blazquez), jdomenech@upvnet.upv.es (J. Domenech).
Technological Forecasting & Social Change xxx (xxxx) xxx–xxx
0040-1625/ © 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).
Please cite this article as: Blazquez, D., Technological Forecasting & Social Change (2017), http://dx.doi.org/10.1016/j.techfore.2017.07.027

Hashem et al., 2015).
The Big Data paradigm also offers many advantages and benefits for
the companies, governments, and the society. Jin et al. (2015) highlight
its potential contribution to national and industrial development, as it
enforces to change and upgrade research methods, promotes and makes
it easy to conduct interdisciplinary research, helps to nowcast the
present and to forecast the future more precisely. In this vein, first Big
Data architectures designed for specific fields are being proposed in
order to surpass the previously mentioned challenges and make the
most of the data available with the aim of nowcasting and forecasting
variables of interest.
However, no specific architecture for social and economic fore-
casting has been proposed yet. This emerges as a necessity, in the one
hand, because of the particular nature of socio-economic data, which
have important components of uncertainty and human behavior that
are particularly complex to model; and, in the other hand, because of
the great benefits that can be derived from the use of Big Data to
forecast economic and social changes. For instance, Big Data ap-
proaches have been proved to improve predictions of economic in-
dicators such as the unemployment level (Vicente et al., 2015), help
managers detect market trends so that they can anticipate opportu-
nities, and also help policy-makers monitor faster and more precisely
the effects of a wide range of policies and public grants (Blazquez and
Domenech, 2017).
In this context, this paper aims to i) establish a framework about the
new and potentially useful available sources of socio-economic data and
new methods devoted to deal with these data, ii) propose a new data
lifecycle model that encompasses all the processes related to working
with Big Data, and iii) propose an architecture for a Big Data system
able to integrate, process and analyze data from different sources with
the objective to forecast economic and social changes.
The remainder of the paper is organized as follows: Section 2 re-
views the Big Data architectures proposed in the literature; Section 3
compiles the new socio-economic data sources emerged in the Digital
Era and proposes a classification of them; Section 4 reviews the new
methods and analytics designed to deal with Big Data and establishes a
taxonomy of these methods; Section 5 depicts the data lifecycle on
which the proposed Big Data architecture is based; Section 6 proposes a
Big Data architecture for nowcasting social and economic variables,
explaining its different modules; finally, Section 7 draws some con-
cluding remarks.
2. Related work
Since the advent of the concept of “Big Data” two decades ago, some
architectures to manage and analyze such data in different fields have
been proposed, having their technical roots in distributed computing
paradigms such as grid computing (Berman et al., 2003). However, the
current data explosion also referred to as “Data Big Bang” (Pesenson
et al., 2010) in which there is a daily generation of vast quantities of
data from a variety of formats and sources, is revealing the fullest
meaning of “Big Data”.
The particular properties and challenges that the current Big Data
context opens require specific architectures for information systems
particularly designed to retrieve, process, analyze and store such vo-
lume and variety of data. Therefore, we are living the constant births of
new technologies conceived to be useful in this context such as, to
mention some, cloud and exascale computing (Bahrami and Singhal,
2014; Reed and Dongarra, 2015). Given this recent technological and
data revolution, research in this topic is in its early stage (Chen et al.,
2014). In this section, we review the novel and incipient research works
that develop general frameworks and specific architectures for adopting
the Big Data approach in different fields from the point of view of data
analytics applications.
Pääkkönen and Pakkala (2015) proposed a reference architecture
for Big Data systems based on the analysis of some implementation
cases. This work describes a number of functionalities expected to be
considered when designing a Big Data architecture for a specific
knowledge field, business or industrial process. These include: Data
sources, data extraction, data loading and preprocessing, data proces-
sing, data analysis, data transformation, interfacing and visualization,
data storage and model specification. Besides that, Assunção et al.
(2015) reflected on some components that should be present in any Big
Data architecture by depicting the four most common phases within a
Big Data analytics workflow: Data sources, data management (including
tasks such as preprocessing and filtering), modelling, and result analysis
and visualization. This scheme was put in relation to cloud computing,
whose potential and benefits for storing huge amounts of data and
performing powerful calculus are positioning it as a desirable tech-
nology to be included in the design of a Big Data architecture. Con-
cretely, the role of cloud computing as part of a Big Data system has
been explored by Hashem et al. (2015).
About architectures for specific domains, Zhang et al. (2017) pro-
posed a Big Data analytics architecture with the aim of exploiting in-
dustrial data to achieve cleaner production processes and optimize the
product lifecycle management. This architecture works in four main
stages: in stage 1, services of product lifecycle management, such as
design improvement, are applied; in stage 2, the architecture acquires
and integrates Big Data from different industrial sources, such as sen-
sors; in stage 3, Big Data is processed and stored depending on their
structure; finally, in stage 4, Big Data mining and knowledge discovery
is conducted by means of four layers: the data layer (mixing data), the
method layer (data extraction), the result layer (data mining) and the
application layer (meeting the demands of the enterprise). Results from
last stage fill the ERP systems and are used along with decision support
systems to improve product-related services and give feedback in all
product lifecycle stages.
In the domain of healthcare, a complete and specific Big Data
analytics architecture was developed by Wang et al. (2016a). This ar-
chitecture was based on the experiences about best practices in im-
plementing Big Data systems in the industry, and was composed of five
major layers: first, the data layer, which includes the data sources to be
used for supporting operations and problem solving; second, the data
aggregation layer, which is in charge of acquiring, transforming and
storing data; third, the analytics layer, which is in charge of processing
and analyzing data; fourth, the information exploration layer, which
works by generating outputs for clinical decision support, such as real-
time monitoring of potential medical risks; last, the data governance
layer, which is in charge of managing business data throughout its
entire lifecycle by applying the proper standards and policies of security
and privacy. This layer is particularly necessary in this case given the
sensibility of clinical data.
The review of these architectures evidenced some common modules
or functionalities. After homogenizing the different names for modules
very similar responsibilities, and considering their sequence in the
process, they can be summarized as follows: first, a data module, which
includes different sources of data with different formats; second, a data
preprocessing module, which includes data extraction, integration and
transformation; third, a data analytics module, which includes model-
ling and analysis techniques for knowledge discovery; and fourth, a
results and visualization module, which includes tools for representing
the results in a way useful for the firm or organization.
However, there are other functionalities whose location within the
Big Data architecture is not homogeneous across the different propo-
sals. For instance, the data storage responsibilities, which are basic for
enabling data reuse and bringing access to previous results, have been
included in a variety of places, ranging from being included in the data
module (Assunção et al., 2015) or the preprocessing module (Wang
et al., 2016a; Zhang et al., 2017), to being a macro-functionality present
in each module of the architecture (Pääkkönen and Pakkala, 2015). The
last approach is better reflecting the nature and complexity of Big Data
analysis, given that not only the original data requires storage, but also
D. Blazquez, J. Domenech Technological Forecasting & Social Change xxx (xxxx) xxx–xxx
2

the integrated data, processed data, and the results derived from data
analytics.
Other functionalities whose consideration in the literature has been
divergent are those related to data governance, which is concerned to
preserve privacy, security and assure the accomplishment of data-re-
lated regulations. Despite its importance, data governance was only
considered by Wang et al. (2016a). As long as the 5Vs model expands,
data governance is expected to gain relevance and become a require-
ment in the design of any Big Data architecture.
For the case of Big Data for social or economic domains, no specific
architecture has been proposed yet in the literature. Given their parti-
cular characteristics and increasing potential for detecting and mon-
itoring behaviors and trends, which is basic to anticipate events, design
better action plans and make more informed decisions, an architecture
specifically devoted to treat these data emerges as necessary. Thus, this
work proposes a Big Data architecture designed for nowcasting and
forecasting social and economic changes. This proposal aims to help
business implement the most appropriate architecture for their decision
making needs, make the most of the data available and assure that it is
treated according to the ethic and legal standards.
3. Non-traditional sources of social and economic data
The digital footprint left by individuals has caused an exponential
growth of the data sources available for social and economic analyses,
which broadens the possibilities for conducting socio-economic studies
beyond traditional data sources, namely surveys and official records.
Although the reasons why these new data are generated are numerous,
the way they are generated has important ethical and legal implica-
tions. For instance, personal data in a purchase order cannot be used for
the same purposes as the data from a public profile in Twitter. To some
extent, the usage of the data is limited by how they are generated. This
fact motivated us to review and classify the newborn non-traditional
sources of social and economic data according to the purpose of the user
generating the data, as Fig. 1 shows.
The first level in the taxonomy includes five categories: i) purpose of
searching for information; ii) purpose of conducting a transaction,
which could be of a financial or non-financial nature; iii) purpose of
disseminating information; iv) purpose of doing a social interaction;
and v) not a deliberate purpose. The first four categories correspond to
an active generation of data, while the last correspond to an inactive
generation: that is, data is not intentionally generated as a result of a
particular purpose, but just derived from the own use of any device (PC,
smartphone, tablet…) with any of the purposes explained above. Data
that fall in this category have been divided in three types: usage data,
location data and personal data. A brief description of each purpose
from which data is generated and examples of sources involved in each
data generation process is shown in Table 1.
The majority of non-traditional sources of social and economic data
mentioned above needs the Internet for working. Indeed, the increasing
penetration and importance of the Internet in almost every social and
economic activity has positioned it as a basic means for the generation
of such kind of data.
3.1. The Internet as basic means for generating socio-economic data
The “Data Big Bang” originated in the Internet, which unstoppable
expands, is transforming the way of interacting in the economic and
social framework. Myriad individuals, companies and public organisms
search, post and generate tons of information daily through the
Internet. These online activities leave behind a digital footprint that can
be tracked and, if treated with the proper Big Data architecture, could
help to describe their behavior, decisions and intentions, and thus, to
monitor key economic and social changes and trends. Indeed, recent
research highlighted the increasing role of the Internet as a provider of
data for explaining, modelling, nowcasting and forecasting social
behaviors (Askitas and Zimmermann, 2015).
3.1.1. Google Trends: the power of search engines
Google Trends (GT) is an Internet-based facility, released on May
2006, which provides up-to-date reports on the volume of search
queries on a specific keyword or text, with historic searches available
since January 2004. It captures how the demand of information under
certain topics varies over time, providing useful data to detect emerging
trends and underlying interests and concerns of society. The use of GT
data to nowcast social and economic (particularly macroeconomic)
variables was introduced by Choi and Varian (2009a,b), who showed
that some search categories in the Google search engine helped to
predict car and home sales, incoming tourists or unemployment claims.
Afterwards, various studies in different countries focused on improving
unemployment-related variables' forecasts by using GT data, obtaining
successful results (Askitas and Zimmermann, 2009; McLaren and
Shanbhogue, 2011; Fondeur and Karamé, 2013; Vicente et al., 2015).
The aggregate consumer behavior in different sectors has also been
successfully predicted with GT data. For instance, using GT data as
predictors has been proved to improve forecasts of tourist inflows
(Artola et al., 2015; Bangwayo-Skeete and Skeete, 2015), of trading
decisions and transaction volumes on the stock market (Preis et al.,
2013; Moat et al., 2014), of private purchases of different goods and
services (Vosen and Schmidt, 2011) or of cinema admissions (Hand and
Judge, 2012). Recently, GT data have proven to be useful for fore-
casting political inquiries' results (Mavragani and Tsagarakis, 2016).
However, elections results and topics with such components of opinion
and ideology have been particularly studied through data from sites
focused on social interaction, as are Social Networking Sites (SNS) such
as Facebook and Twitter and opinion platforms such as Ciao.
3.1.2. Social Networking Sites and blogs
SNS are online places specifically addressed to encourage users
express their feelings and opinions about any kind of topic. Therefore,
the information they contain is to some extent a reflection of what
happens in society. Indeed, the term “Social Big Data” is becoming
popular to refer to data generated by SNS and blogs (Bello-Orgaz et al.,
2016). For that reason, more attention is being paid to SNS as sources of
data potentially useful in forecasting social variables.
Among SNS, the microblogging service Twitter is one of the most
popular, with 332 million users who are active monthly and send on
average more than 500 million tweets per day. This huge amount of
“user-generated” information, though implies some issues, weaknesses
and challenges that require further research (Gayo-Avello, 2013;
Schoen et al., 2013), could help to predict both present and future so-
cial and economic events, as verified in different works. For instance,
tweets' contents have helped to describe political preferences and
forecast elections results (Tumasjan et al., 2011; Kim and Park, 2012;
Ceron et al., 2014), to predict stock market movements (Bollen et al.,
2011), to forecast box office in the motion pictures industry (Kim et al.,
2015; Gaikar et al., 2015) or to monitor the public opinion on new
policies (Ceron and Negri, 2016).
Facebook, which is the third most visited site worldwide1
with
1,650 million active users, doubtlessly also represents a source of
powerful data for analyzing social and economic behaviors. However,
given that its contents are more heterogeneous and user-adjustable,
they are also more difficult to retrieve and analyze. Notwithstanding
this, incipient studies have shown the ability of Facebook data to de-
termine consuming profiles, which are useful for marketing purposes
(Arrigo et al., 2016), and to predict election results and the population's
political orientation (Cameron et al., 2016; David et al., 2016).
Other principal SNS are LinkedIn, Youtube, Instagram, Google+,
Tumblr and Flickr (Bello-Orgaz et al., 2016). They are also rich sources
1
alexa.com.
3

of social and economic data, which could eventually be used to find
changes in the unemployment patterns or detect what entertainment
activities people prefer, among other topics (Russell, 2013). However,
the diverse and complex formats of the information provided, along
with the newness in some of these SNS, makes them remained almost
unexplored. It should be noted that blogs are also important generators
of “Social Big Data”, though research in relating blogs' data to fore-
casting is also in its early stage. The pioneer work of Liu et al. (2007)
examined the usefulness of opinions and sentiments extracted from
blogs to forecast sales performance, while more recently Saleiro et al.
(2015) combined data from news, blogs and SNS to track political
opinion in real-time. Nevertheless, these sources are not without lim-
itations. It is common that they are biased towards one segment of the
population, e.g., young people, and English language, e.g., blogs in non-
English language link more frequently English content than the other
way round (Thelwall, 2007). Thus, some correcting measures should be
considered before generalization (Gayo-Avello, 2012).
3.1.3. Websites and apps: transactional, opinion platforms and information
diffusion
In the Digital Era, firms generally establish their official public
image on the Internet by implementing corporate websites. Through
Fig. 1. Taxonomy of non-traditional sources of social and economic data.
Table 1
Classification of sources of socio-economic Big Data.
User's purpose Description Examples of sources
Information search The user aims to find information about a topic of his interest. Data is actively generated Search engines, Google Trends
Transactions The user interacts with an individual and/or machine to achieve an agreement in which
the user demands and obtains a product or service in exchange for a financial or non-
financial compensation. Data is actively generated.
• Financial transactions Event in which the user makes a payment to obtain a product or service E-banking, e-commerce, urban sensors (tolls, credit card
readers, retail scanners, public transport card readers)
• Non-financial
transactions
Event in which the user provides the counterpart with required information to obtain a
product or service
E-government, e-recruiting
Information diffusion The user aims to spread information or knowledge. This includes marketing purposes, in
order to establish a public image of the user or the agent he represents. Data is actively
generated.
Corporate websites, apps, Wiki pages
Social interaction The user wants to share information, opinions and ideas with other users. Data is
actively generated.
Social Networking Sites, opinion platforms, blogs
Non-deliberate The user does not pursue to generate data with his/her action, but data are generated by
the use of some means. Data is passively generated as a result of any other user action.
• Usage The simple fact of using any device generates data related to how, when and where an
action has been done.
Web cookies, Internet Protocol, Sensors for self-tracking
• Location The use of mobile phones generates data particularly related to the position of the user. GPS, GSM, Call Detail Records, Bluetooth, WiFi Points
• Personal Personal data (age, sex, etc.) is generated consciously (e.g. filling a form to complete a
purchase) or unconsciously (e.g. data about the type of information we look for is used
to infer our incomes) as a consequence of using any device or tool to achieve a purpose.
Forms, profiles, type of searches or purchases
4

these sites, companies inform about their products, services, organiza-
tional structure and intentions, such as exporting and opening a branch
office abroad. Corporate websites encompass all kind of websites im-
plemented by firms in relation to their economic activity, ranging from
websites used only to give information about the firm, to transactional
websites devoted not only to provide information but also to offer on-
line services (e-commerce, e-banking…), about which users are some-
times allowed to give their opinion in the website itself. That is, cor-
porate websites may present three different functionalities: spreading
information about firms (related to establishing a public image), con-
ducting transactions (e-business processes) and facilitating opinion
sharing (electronic word-of-mouth (eWOM) booster).
It is remarkable that websites have a complex structure which differ
from one case to another, so that standardizing the retrieval and ana-
lysis of their information requires from a specific Big Data architecture.
That difficulty has contributed to corporate websites being an almost
unexplored source of data. However, their public, updated and “busi-
ness generated” nature makes them potential sources of economic data.
Moreover, business characteristics could emerge on the web and be
monitored by massively analyzing corporate websites, as recent re-
search shows.
Applying Big Data approaches (particularly web data mining and
machine learning) to the “spreading information” functionality of cor-
porate websites, firms' sales growth and business activities such as
strategies of technology adoption, innovation and R & D have been
successfully detected (Arora et al., 2016, 2013; Gök et al., 2015; Li
et al., 2016). In addition, by using a specifically designed web data
mining system for analyzing corporate websites (Domenech et al.,
2012) the export orientation of firms has also been successfully de-
tected (Blazquez and Domenech, 2017). In addition, there exist other
type of websites created with the specific aim of spreading information,
such as are Wiki pages, from which Wikipedia is the most important
representative nowadays with more than 730 million unique visitors
monthly (Wikimedia Foundation, 2017). Its penetration in the society
along with its collaborative nature have positioned it as a potential
source of social and behavioral data. Concretely, Wikipedia page views,
edits and contents have already proven to be useful for socio-economic
forecasting. Incipient research works have successfully used Wikipedia
data to better forecast stock market movements (Moat et al., 2014) and
tourism demand (Alis et al., 2015; Khadivi and Ramakrishnan, 2016).
This kind of studies aim to create new indicators in advance or to
complement those used in current official statistics.
The prominent role of the Internet in today's economy and society
has promoted the emergence of e-business services, which firms can use
to sell their products and do transactions in an online base with cus-
tomers (E-commerce), recruit candidates (E-recruiting) or offer their
services online (e.g. E-banking). E-business may even go a step further
and represent not only a complementary tool for firms (e.g., selling
locally and online), but a new type of business model characterized by
operating just online. Many of these sites offer users the chance to post
their opinions and do reviews on the product or service acquired, which
may range from any manufacture to a hotel stay or an experience in a
restaurant. This web feature is generally known as opinion platform
(even a website can be designed just to act as opinion platform), which
is used to bring together online communities of users, whose opinions
are basic information for social science research.
One of the most important e-commerce and opinion platform
worldwide is Amazon. It is one of the biggest online retailers, with more
than 300 million active customers' accounts. This website provides
customers' reviews and opinions on millions of products and services,
being therefore a source of data potentially useful to detect consumer
preferences or predict sales. For instance, the forecast of consumer
product demands in Amazon.com has been significantly improved by
using the textual contents of consumer reviews (Chong et al., 2015;
Schneider and Gupta, 2016). Another noteworthy topic for managers is
to detect the so-called “influencers” in consumer-opinion platforms,
given that their comments may influence other consumers' purchase
behavior and, thus, detecting and monitoring them is essential. For
instance, Arenas-Márquez et al. (2014) successfully identified influen-
cers in Ciao.com by retrieving and analyzing characteristics of the
product reviews such as the rating received by other users.
Other sites that provide potentially useful data for detecting social
and economic trends are, for instance, eBay.com, whose information
has been helpful to explain price differences among remanufactured,
used and new items (Frota Neto et al., 2016), TripAdvisor.com, which
has been successfully used to detect tourist preferences thus helping
hotel managers to adapt their offers (Li et al., 2015), and Monster.com,
which organizes the available job offers and helps to track changes in
job search (Edelman, 2012).
When using opinion platforms as sources for social and economic
analyses, limitations related to the veracity of the contents must be
considered. Sellers and marketers may have the temptation to generate
fake consumer reviews to influence in the consumer decision (Malbon,
2013). In this context, some techniques for detecting such manipula-
tions could be applied to alleviate this limitation (Hu et al., 2012).
Apps provide access to information and services that may or may
not be offered by other means, such as websites. Since the use of apps is
becoming widespread in the daily activities of individuals and organi-
zations, they have become a source of data with great potential for
forecasting social and economic topics. Although accessing data gen-
erated by them is currently a difficult task, some incipient research
works are appearing. To date, apps data logs have been proved to be
successful for forecasting users' intentions to use a specific app, auto-
matically forecasting depression (Wang et al., 2016b; Suhara et al.,
2017) or helping to detect mobility patterns as reviewed by Pan and
Yang (2016).
3.2. Urban and mobile sensors
Ubiquitous computing is one of the technological areas that has
experimented the greatest development in the context of the Digital
Era. Its advances have resulted in the generation of wireless, incon-
spicuous and inexpensive sensors to gather information on our ev-
eryday life activities (Krishnan and Cook, 2014). Specifically, urban
sensors and mobile embedded sensors are potential generators of social
and economic data.
Among urban sensors, one of the most widespread and used
worldwide is the credit card reader. Credit card transactions are re-
corded and provide data potentially useful for firms to detect and
predict, for instance, personal bankruptcy (Xiong et al., 2013), frau-
dulent purchases in online stores (Van Vlasselaer et al., 2015) and de-
fault and repayment, which in the context of credit card companies is
useful for defining marketing strategies (Einav and Levin, 2014).
Retail scanners are also very extended, and their function is to re-
cord the characteristics of customers' everyday purchases. These data
has proven to be useful for forecasting consumer behaviors, sales and
prices, as recent research shows. For instance, Dey et al. (2014) suc-
cessfully used retail level scanner data to model market trends, prices
and sales in the industry of catfish products, suggesting a particular
competition strategy based on the results obtained. Another study, fo-
cused on explaining human behavior, employed weekly scanner data to
detect consumer boycotts in response to an international conflict
(Pandya and Venkatesan, 2016).
A pioneer study by Askitas and Zimmermann (2013) successfully
used data from tolls to nowcast business cycles, creating a Toll Index
that represents a technological, innovation-driven economic telemetry.
Other sensor networks that provide useful data for forecasting a
manifold of socio-economic variables are smart grid, WiFi access points
and public transport card readers, among others (Kitchin, 2014; Chou
and Ngo, 2016).
Some sensors embedded in mobile phones are also potential sources
of social data: GSM, GPS, Bluetooth, accelerometer or sensors for
5

connecting to the telephonic network through Base Transceiver Stations
(which produce the so-called “Call Detail Records”, with information
regarding all call-related activities, such as sending SMS and phoning,
conducted by mobile phone users in the network). These sensors gen-
erate data related to the user location that have been successfully used
to study social behaviors, preferences and mobility patterns. Properly
treated, these data can contribute to better understand in which way
human mobility affects well-being and human behaviors at the micro
level, and social organization and change at the macro level (Williams
et al., 2015).
Concretely, data from such sensors have been useful for detecting
places of interest, that is, places where people go and stay for a while
(Montoliu et al., 2013), and for detecting personality traits, which
companies may use to personalize their services (Chittaranjan et al.,
2013). Moreover, Laurila et al. (2013) summarized different human
behaviors analyzed to date with such mobile embedded sensors data,
including: mobility patterns and their relation with the weather, the
perceived level of safeness and intimacy of a given location, the relation
among moves from individuals and from their friends and acquain-
tances, and the transition between spatial habitats. Other recent ap-
plications of mobile phones' data in relation to mobility are recreating
and drawing maps of population distribution (Deville et al., 2014;
Graells-Garrido et al., 2016) and detecting anomalous behavioral pat-
terns associated to emergency (e.g. earthquakes) and non-emergency
(e.g. holidays) events (Dobra et al., 2015).
4. Non-traditional methods for processing social and economic
data
Data obtained from non-traditional socio-economic sources are
generally large, heterogeneous and unstructured or semi-structured.
These characteristics imply a number of challenges when it comes to
retrieving, processing, analyzing and storing data. Accordingly,
methods and techniques related to machine learning and Big Data are
being developed. Many of such methods have been widely applied in
other knowledge fields such as engineering, medicine and biostatistics.
Despite their potential for treating socio-economic data, their applica-
tion in this field is still at an early stage (Varian, 2014).
This section enumerates and describes the most relevant methods
for treating socio-economic data from a Big Data approach, with the
objective of providing a framework. The reviewed techniques are
summarized and classified in a taxonomy illustrated in Fig. 2.
4.1. Methods for structuring data
Big Data sources can be classified as structured (tabular data), semi-
structured (data with machine-readable tags that do not follow a strict
standard) or unstructured (data that lacks from any scheme allowing
machines to understand them, e.g. a video). Since analysis algorithms
require some structure to interpret the data and given that about 95% of
Big Data is unstructured (Gandomi and Haider, 2015), the process of
structuring the information is basic. This includes transforming the data
into an organized set, with clearly defined variables and the relations
among them identified. Below, some of the most common methods for
structuring data with applications to social and economic analyses are
surveyed.
Almost any source of data, and particularly the Internet, is plenty of
human generated text that requires proper retrieval and processing. To
exploit the full potential of text in databases, specific techniques for
processing natural language are required. Natural Language Processing
(NLP) is a research area focused on exploring how computers can be
used to understand and shape natural language text so that it can be
useful for different applications (Chowdhury, 2005). NLP is in itself a
computational method that comprehends a series of techniques that
provide an easy interface for information retrieval systems and, at the
same time, to structure texts in different ways so that the underlying
information can be more easily extracted. Some interesting NLP tech-
niques for social analysis are Sentiment Analysis (also referred to as
Opinion Mining), Latent Semantic Analysis (LSA), Latent Dirichlet Al-
location (LDA), TF-IDF (Term Frequency - Inverse Document Fre-
quency) and Word Embeddings. Liu (2012), Evangelopoulos et al.
(2012), Blei et al. (2003), Moro et al. (2015), Armentano et al. (2014)
and Rudolph et al. (2016), among others, provide some reference for
these methods.
Linking records from the same user (or entity) across different data
sources is also an important challenge for analyzing social and eco-
nomic information. Data Matching (which is also commonly known as
Record Linkage or Entity Resolution) is a computational process used to
identify, match and merge records from several databases that corre-
spond to the same entities. A special case of data matching is dedu-
plication, which consists in the identification and matching of records
about the same entities within just one database (this step is crucial in
data cleaning). Matched data are becoming more important because
they may contain information impossible to obtain by means of other
sources or processes. This technique is a complex process encompassing
five steps, from data cleaning and standardization to data quality and
completeness measuring. For a detailed description, see the work by
Vatsalan et al. (2013).
4.2. Methods for modelling data
Modelling data (and their relationships) is the main process in a Big
Data analysis. This includes reducing the dimensionality of data sets,
applying modelling techniques to data and obtaining outcomes.
Depending on the type of data available and the objective of the ana-
lysis, two different paradigms for modelling data may be applied:
Supervised Learning and Unsupervised Learning (Hastie et al., 2013).
On the one hand, Supervised Learning refers to problems in which
each observation in a data set has inputs (also referred to as in-
dependent variables, features or predictors) and outputs (also referred
to as targets, responses or dependent variables), and the main goal is to
use inputs in order to infer the values of outputs. These problems can be
further categorized as classification problems, in which outputs are
expressed as categories, or as regression problems, in which outputs are
expressed in a continuous space. On the other hand, Unsupervised
Learning refers to problems in which each observation has some inputs
but no outputs, and the main goal is to find the relationships or struc-
ture among inputs. These problems can be further categorized into
clustering problems, in which the goal is to discover groupings in the
data, and association problems, in which the objective is to find rules
that describe the behavior of part of the data.
Depending on the learning paradigm, different machine learning
techniques can be applied. For nowcasting and forecasting applications,
supervised methods are generally employed. The most common su-
pervised machine learning techniques successfully applied in other
disciplines, such as medicine and engineering, and that are potentially
useful for the social sciences, are enumerated below.
Linear and logistic regressions are two useful machine learning
techniques widely applied by economists and social scientists.
However, alternative methods to regressions have been developed and
demonstrated to perform as well as or better when using big data sets
(Varian, 2014). For instance, Decision Trees, which are a type of pre-
dictive models that can be used to represent both classifiers and re-
gression models; Support Vector Network (Cortes and Vapnik, 1995),
more commonly known as Support Vector Machine (SVM), which is a
learning machine for two-group classification; Artificial Neural Net-
works (ANN), which are two-stage regression or classification models
able to identify non-linear relations among a set of input variables, and
generate forecasts about the variable under study by modelling and
weighting those relations (Hastie et al., 2013); and Deep Learning
methods, which develop a layered and hierarchical architecture where
higher-level (more abstract) features are obtained by transforming
6

lower-level (less abstract) features. For classification, higher-level fea-
tures highlight aspects of the raw input that are relevant for dis-
crimination. Deep Learning methods can deal with huge quantities of
unstructured data, reason why they are positioning as a promising tool
in Big Data analysis (LeCun et al., 2015; Najafabadi et al., 2015). ANN
and Deep Learning are special cases, given that their learning algo-
rithms can be either supervised or unsupervised.
In addition, there exist a group of techniques which are focused on
improving the performance of the previously described ones, and that
are starting to be known as “Ensemble algorithms”. Some of these al-
gorithms work by adding randomness to data, which is a useful pro-
cedure to deal with overfitting. These techniques include the Bootstrap,
Bagging, Boosting and Random Forests (Varian, 2014).
Regularization methods are another group of supervised learning
techniques, whose objective is to obtain sparse solutions and that, due
to the increased amount of information available, have been increas-
ingly studied in recent years by the scientific community (Friedman
et al., 2010). These methods can be applied to a number of supervised
learning techniques, from regressions to support vector machines, and
include, to mention some examples: the Least Absolute Shrinkage and
Selection Operator (LASSO), which was one of the first regularization
methods (Tibshirani, 1996); the regularization for support vector ma-
chines (Hastie et al., 2004); the Elastic Net, which is a mixture of the
LASSO and Ridge Regression (Zou and Hastie, 2005); and a regular-
ization scheme for neural networks, aimed at improving the classifi-
cation margin (Ludwig et al., 2014).
Finally, Bayesian Statistics constitute an alternative approach to
frequentist statistics (as are the methods describe above) in both terms
of decision theory and inference. Though their potential in the social
sciences and economy was pointed out almost 40 years ago (Harsanyi,
1978), the complex numerical integrations needed made them remain
unused. However, the recent advances in computation methods have
made it possible to easily apply Bayesian methods (Congdon, 2007).
To mention some, Bayesian Model Averaging (BMA) is a multi-
modelling method that is starting to be applied to linear regression and
generalized linear models (Ley and Steel, 2012). Naive Bayes, which is
Fig. 2. Taxonomy of non-traditional methods to treat social and economic data.
7

a machine learning tool for classification whose popularity is starting to
increase due to its simplicity for being implemented, being fast and
computationally efficient, and obtaining high classification accuracy,
especially for Big Data (Wu et al., 2015). Also, the Spike-and-Slab Re-
gression, which is a variable selection method for linear regression
models (Varian, 2014). Besides this, the Bayesian Structural Time Series
(BSTS) technique is devoted to treating panel or longitudinal data,
which are very common in the social sciences. This is a method for
variable selection and time series forecasting and nowcasting, used as
an alternative to traditional time series analysis methods such as Au-
toregressive (AR) and Moving Average (MA) models.
4.3. Methods for assessing models' performance and robustness
A basic objective in any data analysis focused on forecasting is to
obtain a robust model with the best out-of-sample predictive precision
possible. In this subsection, a brief review on techniques for improving
the performance of forecasting and nowcasting models is provided.
Assessing the performance and robustness of predictive models is
essential to determine their validity and applicability, and the quality of
the predictions. In this case, performance refers to how well a model fits
the data and how accurate it is, while robustness refers to how well a
model works on alternate data, that is, on data which is different from
that used to build the model. If a model has a good performance, then it
is capable of detecting the characteristics in a data set and providing
highly accurate predictions. Moreover, if it is robust, then the predic-
tions obtained could generalize and so the model is valid and useful
with new data. The goal in any Big Data analysis is to build models that
simultaneously are robust and provide accurate outputs: this is the only
path to use them as reliable tools for forecasting and nowcasting whose
results can be used for decision-making.
To compare and select different kind of models depending on how
well they fit to data and how complex they are, there exist uncountable
classically applied tests such as Nagelkerke's R2
, Hosmer-Lemeshow,
Mallows' Cp, Akaike Information Criterion (AIC), Bayesian Information
Criterion (BIC), Deviance and Log-Likelihood, among others. Although
these tests and indices provide useful information about model per-
formance, they were not conceived for treating the huge amount of
complex data with which we work nowadays. The particular char-
acteristics and issues of Big Data (size, bias, imbalanced sets, complex
relations…) make necessary to complement classical tests with more
recently developed techniques that are capable to better deal with these
issues (Varian, 2014).
First of all, to ensure that the predictions obtained are robust it is
recommended to build the models by conducting a holdout process in
which the initial sample is split into two subsets: the training set and the
test set. The former is used to train the model, and generally includes
about 75% to 90% of the initial observations, while the latter is used to
evaluate its predictive performance and includes the remaining per-
centage of observations. Even if data is large enough, it may be divided
in three sets: a train set (the largest), a validation set and a test set. This
method ensures that the predictions obtained are robust, so that they
can be generalized to an independent data set. Another approach with
the same objective is K-Fold Cross-Validation. In this method, data are
split into K parts of equal size and the model is fitted K times, where K-1
parts are used to train the model and the remaining is used to test its
predictive performance. Finally, the K estimates of the prediction error
are combined. In case each part includes just one observation, then the
process is called Leave-one-out Cross Validation (Hastie et al., 2013).
For big data sets, the first method is recommended.
In addition, for properly training classifiers, at least the train set
should be balanced, because this way the model is built to successfully
detect each of the categories equally. Otherwise the learning process
could be endangered (Menardi and Torelli, 2014). A sample is balanced
when each of the categories of the response variable is present in the
same proportion. To balance an unbalanced data set, solutions such as
oversampling, undersampling, synthetic sampling and kernel methods
can be applied (He and Garcia, 2009). Unbalanced data sets are
common in social sciences, so it is expected that the use of these pro-
cedures in socio-economic research will expand in the near future.
Moreover, think if what we are trying to predict is if someone is
infected with a disease; then, the best situation would be to obtain a
true negative (individual not infected). That is, not only false positives,
but also false negatives, imply costs. It is important to assign a mone-
tary value to these costs in order to influence the decision making of a
model. This process is known as “Cost-sensitive analysis” (Sun et al.,
2007). It makes use of the Cost Matrix, which reflects the costs and
benefits associated to each of the four possible outcomes of a classifier.
Providing this information to a classifier, it can be influenced to mini-
mize the most costly errors or to maximize beneficial classifications, so
that we obtain a “weighted accuracy”. Similarly, by using Loss Func-
tions, classifiers are forced to give preference to predictors that help to
predict true probabilities accurately (Witten et al., 2016).
To check the predictive accuracy of classifiers, methods such as the
Lift analysis, Precision-Recall Curves, ROC Curves and the Confusion
Matrix are pertinent, whose fundamentals and applications regarding
the social sciences can be looked at (Fawcett, 2006; Witten et al., 2016).
When the output variable is not categorical, but numerical, other
measures are available, such as the Root Mean Squared Error (RMSE),
the Percentage Error (PE), the Fractional Bias and the Index of Agree-
ment (IA), whose popularity is starting to increase.
5. The data lifecycle
Digital data have many advantages, such as being easy to share,
replicate and recombine, which make them potentially reusable.
Business and researchers can take advantage of this to boost research in
progress and leverage past investments, for instance. However, to ex-
ploit all the benefits of digital data, they must be properly collected,
processed and preserved. Data loss or damage may imply economic
costs as well as lost chances, reason why funder agents (public or pri-
vate) are increasingly demanding institutions to document and run
data-management plans taking into account the whole lifecycle of data
(Lynch, 2008). For this reason, it is basic to define what phases and
processes form this lifecycle in order to implement robust and flexible
architectures to manage data in the context of the Digital Era.
The data lifecycle is the sequence of stages that data follow from the
moment they enter a system to the moment they are erased from the
system or stored (Simonet et al., 2015). Between the data entrance and
exit or storage, data go through different stages, which may differ de-
pending on the type of data and purpose to achieve as documented in
the compilation of classic data lifecycles (Committee on Earth
Observation Satellites — Working Group on Information Systems and
Services, 2012). The Knowledge Discovery in Databases (KDD) process
was the first proposal of a model to manage digital data (Fayyad et al.,
1996). It refers to the complete (non-cyclical) process of extracting
knowledge from data, and includes five main stages: data selection,
data preprocessing, data transformation, data mining and data inter-
pretation. As databases started to exponentially grow in size and
complexity, the necessity of a wider scheme to appropriately manage
these data was highlighted, especially by the industry. This derived into
the development of the Cross-Industry Standard Process for Data
Mining (CRISP-DM process), which is an expanded implementation of
the KDD process that introduced the management of digital data as a
cycle (Chapman et al., 2000). It comprises six stages: business under-
standing, data understanding, data preparation, modelling, evaluation
and deployment. If both are compared, the first and last stages of
CRISP-DM process are new with respect to the KDD process, while the
“data understanding” stage of CRISP-DM is similar to the “data pre-
processing ” and “data transformation” stages of KDD.
The next approach to data management within a digital environ-
ment that the scientific and industrial community focused on, and to
8

which most research efforts have been paid since these days, was called
itself the “data lifecycle”. The Data Documentation Initiative Alliance
(DDI Alliance) was one of the first voices to focus their efforts on this
idea (DDI Alliance, 2008). It proposed a data lifecycle including the
following five stages: first, discovery and planning; second, initial data
collection; third, final data preparation and analysis; fourth, publication
and sharing; and last, long-term management. This departing point
considers from planning the project (what is being studied, what data
are needed and how they are going to be treated, etc.) to determining
how to store and preserve data in the long-term. With respect to KDD
and CRISP-DM processes, this is a more extensive approach that in-
cludes important concepts within digital data such as sharing and long-
term management.
Afterwards, Corti et al. (2014) described the phases and activities
typically undertaken within the research data lifecycle. These phases,
each of which included a number of specific activities, are the fol-
lowing: discovery and planning, data collection, data processing and
analysis, publishing and sharing, long-term management and reusing
data. This proposal extends the initial one by DDI Alliance to include an
additional stage at the end of the cycle devoted to data reuse. A more
exhaustive data lifecycle to date was proposed by Rüegg et al. (2014),
who included up to eight stages: the first four stages (planning, data
collection, data quality control, and analysis) correspond to managing
data in a traditional project which is new (no previously results or data
exist). If the project relies on existing data (referred to as “data reuse”),
then it follows the third first stages and continues with additional data
discovery, data integration, and finally, the analysis.
While Corti et al. (2014) consider data reuse as a step itself, Rüegg
et al. (2014) reference data reuse as a type of project in which existing
data are used, including some steps within this lifecycle. The context of
economic and social analyses makes it more appropriate to consider
data reuse as a step itself, given that as a project that started from
scratch develops and data are obtained and exploited, these data may
be reused many times in the same project with different purposes. That
is, the view that a project is new or departs from data seems excessively
static for economic and social nowcasting purposes. Additionally, to
complete each of the data lifecycles, this work includes two more steps:
data documentation and data archiving in a public repository, which
we consider basic for preserving and publishing data.
The review of these works of reference allowed us to integrate and
fully describe the different stages of a full data lifecycle in the context of
economic and social analyses. Its aim is to standardize the concept of
data lifecycle and serve as framework when it comes to designing a
proper data management architecture in this context. Our proposal for a
data lifecycle includes nine stages, as reflected in Fig. 3. These stages
are described as follows:
1. Study and planning: This first stage consists in designing the re-
search or business project to achieve the desired goals of funders or
managers. Once each phase of the study is defined, it is necessary to
plan what procedures to treat data (collected or generated
throughout the research) will be applied. For instance, this includes
planning what type of data are going to be collected, how and from
which sources, which methods will be used for their processing and
analysis, where will they be stored, and to find out what legal reg-
ulations and privacy issues affect the type of data that is going to be
analyzed, in order to adapt the operating procedures.
2. Data collection: This stage consists in accessing the sources, which
can be internal or external, and collecting initial or raw data.
Depending on the field of knowledge and the data required for de-
veloping the project, activities such as phenomena observation,
experimentation, recording, simulating, scraping and negotiating
with third-party data providers will be part of this stage.
3. Data documentation and quality assurance: This stage consists in
documenting the acquired data and checking their quality. First, the
data acquisition process should be documented by associating data
to metadata. The metadata include information related to the source
of origin, data format, technical details on the retrieval process and
accessing dates, among others, thus enabling their reuse and correct
referencing. Second, data quality and validity should be assured. It
is required to verify the trustworthiness of the data sources as well
as of the own data, to control for any data inconsistencies, such as
unexpected values and typing errors, and to clean and anonymize
data if necessary.
4. Data integration: This stage consists in fusing data obtained from
different data sources with a coherent and homogeneous structure,
which helps to make data traceable and easier to access and ma-
nipulate in successive projects. This includes activities such as es-
tablishing relations among variables of different data sources,
adapting units, translating, and creating a single database with all
the acquired data. Data integration should also incorporate privacy
constraints to avoid disclosing some private information in the in-
tegrated data. This is a major concern because rich integrated data
may facilitate discovering some personal details otherwise anon-
ymous.
5. Data preparation: This stage consists in transforming data so that
they meet the format requirements of the analysis tools and tech-
niques that are going to be applied. This includes activities such as
transcribing, digitizing, interpolating, establishing a tabular format
in the data set and deriving new data by operating with the existing
data.
6. Data analysis: This stage consists in analyzing data, obtaining and
interpreting results, and achieving conclusions. A huge range of
statistical techniques and computational tools are called to be used
in this stage. The final selection of the most appropriate techniques
will depend on the type of data analyzed and research objectives.
The interpretation of the results and conclusions achieved, as well as
the results themselves, are basic inputs for the next stage.
7. Publishing and sharing: This stage consists in publishing results and
conclusions derived from data analysis, or the generated data sets
themselves. The outputs of this stage aim to facilitate the decision-
making process of managers or policy-makers (when data is pre-
sented in reports, for instance), to spread knowledge (if a research
article is published, for instance) and to feed automatic systems of
companies with information of relevance to help the staff make
decisions such as ordering supplies, among many others. Other re-
lated activities in this stage are establishing copyright of data and
results, authoring publications, citing data sources, distributing data
and controlling data access.
8. Data storage and maintenance: This stage consists in archiving and
registering all the data gathered, processed and analyzed, for al-
lowing long-term data preservation, curation and reuse. Actions to
be done may include storing data in specific repositories or com-
putational systems, migrating them to other platforms or mediums,
regularly backing up the data, producing associated metadata, pre-
serving the documentation generated during the whole process,
controlling data security and privacy and erasing data if required by
legal regulations, for instance.
9. Data reuse: This stage consists in reusing data that have been pre-
viously gathered, processed, analyzed and stored. This action can be
originated in a variety of different purposes such as testing new
hypotheses related to the same project for which data were col-
lected, sharing or selling data to companies, conducting new pro-
jects for which existing data can be useful and using data with in-
structive purposes.
6. A Big Data architecture for nowcasting and forecasting social
and economic changes
Our proposal of a Big Data architecture for nowcasting social and
economic changes is presented in Fig. 4. Departing from the approach
of the data lifecycle in the organization, it includes layers and modules
9

to manage the processing and integration of social and economic data,
including the storage, processing policies and publication of results.
This architecture is organized in three layers. The data analysis
layer contains the main processes of generating knowledge from the
input data: from the ingestion of data from multiple sources to the
publication of reports. Together with this layer, there are two other
layers that work as support to the data analysis: The governance layer is
in charge of applying policies and regulations to the whole data life-
cycle, as well as managing the licenses related to the data sets. The
persistence layer deals with the storage and management of data to
make them available to the different modules in the data analysis layer.
6.1. Data analysis layer
The data analysis layer is the part of the architecture that imple-
ments the main processes required to generate knowledge, in form of
reports or predictions, from the different data sources to which the
organization has access. It is composed of six modules that work se-
quentially, from the data reception to the publishing of results.
6.1.1. Data receiving module
This module constitutes the data ingestion point in the architecture,
so that data external to the system are made accessible to the other
modules of the architecture. This way, external data are connected to
the processing stream of the nowcasting system. It is composed of dif-
ferent elements, as many as different data sources are used as input.
Connected data sources can be classified into two main groups:
First, those sources owned by the organization implementing the ar-
chitecture. These sources may include relational databases managing
the daily operation of the company, that is, business-specific informa-
tion, such as sales, customers, purchases, website analytics, and so on.
Own sources also involve data not directly generated by the business
operation, but collected or requested by the organization at variable
frequencies. This includes surveys, market research data, and non-
periodic report data.
The second group of data sources are those external, that is, those
sources which are not controlled by the organization, though they may
contain information relevant to its operation. A wide variety of sources
may be considered as relevant for the organization purposes. For in-
stance, some open data offered by public institutions might provide
some information on the context of the company customers. Similarly,
social and economic data published by the official statistics institutions
have also potential for explaining the context in which individuals
make decisions. Google Trends and social media platforms, such as
Twitter and Facebook, are useful sources for detecting trends and re-
lations between relevant topics. Furthermore, many other websites or
RSS providing product opinions, political comments, product releases,
etc. might be explored to find some other contextual variables that
could complement own data sources.
Since the access to these sources is widely heterogeneous, the ele-
ments of this data receiving module must hide the complexity for ac-
cessing the sources. The access to these sources by the different ele-
ments in the module may be done by means of an Application
Programming Interface (API) when available from the data provider, or
by means of specific software developed for this purpose, e.g., web
scraper.
All elements in this module will receive the data with the format
and structure provided by the origin, which could be incompatible
among them. According to their structure, data can be classified as
structured, semi-structured or unstructured. Structured data includes
information organized at high level, such as in relational databases,
which apart from data, contains a schema with restrictions and rela-
tions. Semi-structured data also have some organization of the in-
formation, although the schema is embedded in the data. That is, its
structure is self-describing, as in XML documents. Unstructured data
provide no structure at all, and can be considered as a collection of
Fig. 3. The data lifecycle within a Big Data
paradigm.
10

elements. This does not mean that each element does not have a
structure, but that the schema has not been described, so additional
techniques to infer some structure should be applied. Text documents
are typical unstructured sources. Data structure is a key factor to suc-
ceed integrating data from different sources, as it is the first step to
establish the links between them. Structured data are usually related to
SQL databases, while NoSQL are more suitable for storing unstructured
and semi-structured data.
The elements in this module could access origins in batch or in
stream. Stream processing is possible when the source allows access
with high bandwidth and low latency conditions, e.g., when accessing
an own relational database. However, when access conditions are not so
favourable, the elements should work in batch, thus requiring persistent
storage for the accessed data. In such event, the type of storage (SQL or
NoSQL) must be consistent with the data source type. In any case, the
data access that this module provides to the remaining modules of the
architecture should be given as in stream processing.
6.1.2. Data preprocessing module
This module departs from the data connections prepared in the data
receiving module. It aims to validate and preprocess data to leave them
ready for integrating different sources. This preprocessing is divided in
three steps.
The first step is to record and document the procedure of data ac-
quisition by attaching metadata to the original source. These metadata
should include information about the data source, the version of the
collector (in the receiving module) used for the retrieval, the schema
with the data structure (if any) and other technical details such as the
codification, format and so on.
The second step is to check the internal validity of each source.
Although structured sources usually keep all observations in the right
format, other sources may be internally inconsistent. Thus, this step
involves checking observations for anomalous values (e.g., text when a
number is expected) and dealing with them, for instance, by marking
them as missing or wrong. This may result in a number of useless ob-
servations, that is, those with an excessive number of missing features,
which may be cleaned up to avoid including noise in the data analysis
process.
The third step is related to the extraction of features and the gen-
eration of new data derived from the original source. At this step, only
derived data at entity level should be created. That is, if the origin
provides rows, only data from each row may be used to generate new
features. If the origin provides documents, only document contents may
be used to generate variables describing the document, for instance, by
applying natural language processing techniques. Examples of entity-
level derived data may include counting the number of words of a
comment, detecting the language and computing the term frequency.
Derived data whose computation requires analyzing several entities
(e.g., computing an average) should be generated in the data analytics
module. When the computational effort to generate new data is high,
the resulting features should be persisted to allow for reusing them in
subsequent data integrations. This involves using a database consistent
with the structure type of the origin.
Fig. 4. Big Data architecture for nowcasting and forecasting social and economic changes.
11

6.1.3. Data integration module
The objective of this module is to merge the different data sources
and provide homogeneous access to all data available to the organiza-
tion. To do so, data integration must deal with five challenges: i) het-
erogeneous sources, whose access was homogenized by the data re-
ceiving module; ii) related data structures whose relation has not been
explicitly established by the sources; iii) a variety of data sizes and
probably inconsistent formats; iv) heterogeneous time frequencies,
ranging from milliseconds to years; and v) heterogeneous geographic
groupings, ranging from detailed GPS coordinates to state or country
level.
To relate data from different sources, it is required to define
schemes that establish the relation among them. For instance, estab-
lishing the relation of a commercial establishment to a region, it will be
possible to relate its sales to the average income of the area in which it
is located. These data usually come from different sources: sales are
accessed through internal sources, while the average income could be
provided by an official statistics institute.
To establish such relations, some linkage techniques and hier-
archical groupings might be applied. Geographic hierarchies are useful
to link records to the most appropriate geographic level, which is not
necessarily the lowest one. For instance, street-level economic situation
may be useful for analyzing housing prices, but it is too specific for a
business whose influence area is wider, e.g., an airline office. Linkage
techniques are required when the same entity does not receive the same
identifier across the different sources. This could happen simply be-
cause it is written in a different language (e.g., the country name), si-
tuation which can be solved with a simple translation; but also because
of lacking of a standardized or public id. In such cases, some analysis to
match the record could help find relations and provide new insights on
the data.
Adapting time frequencies is also included in this category. It is
required to adopt some criteria to generate high frequency data from
lower frequencies and vice versa. Reducing time frequencies may in-
volve computing some summarizing statistics (e.g., average and max-
imum), while increasing time frequencies may involve interpolating
data or selecting the closest value in time.
Once several data sources are integrated and their relations are
established, they could be stored in the persistence layer and feed the
data analysis layer again as a new element in the data receiving module.
This way, it is possible to use these sources as a single one when in-
tegrating with additional sources.
6.1.4. Data preparation module
The organization in which data are stored after the integration may
not be suitable to perform the analysis. This module takes the data as
prepared by the data integration module and transforms them to match
the format expected by the data analytics module. Since each element
in the analytics module may expect data in a different format, data
preparation is also specific to each analytics element.
These transformations may involve grouping some elements or
joining data from different tables to enrich the information about each
entity or individual. This is also the most suitable module to alleviate
the missing data, which may be estimated or interpolated to avoid
losing cases.
A common operation in this module is the pivot transformation.
Storing information as key-value pairs, in which entity features are
spread among many rows, is quite convenient in Big Data environ-
ments. However, this may not be the table format expected by the
analysis software. By applying the pivot transformation, all features
regarding the same entity are arranged in the same row, which is the
data organization commonly required to feed the analysis.
The resulting data after the preparation process should be stored to
provide consistent input to the analysis. After conducting data in-
tegration and depending on the purpose of each particular study, it is
possible to obtain small data sets derived from the initial big data set,
which could be treated with traditional statistical techniques. The sto-
rage in this module must be analysis-driven, unlike the previous mod-
ules, whose storage is source-driven.
6.1.5. Data analytics module
This module applies statistical and machine learning methods to
extract knowledge and make predictions from the data prepared by the
previous module. To do so, descriptive and predictive techniques are
applied. The descriptive analysis could provide some insights on the
characteristics and evolution of the socio-economic variables under
study. Its results will be used in the results publishing module to create
tables and graphics representing the relationship among variables.
Predictive techniques are based on models that help explain, clas-
sify, forecast or nowcast the socio-economic variables under study. To
do so, the models are estimated or trained by using learning methods
and relying on any of the methods described in Section 4 for selecting
the most meaningful variables and improving predictions. The com-
puting-intensive nature of these techniques makes it more challenging
to deal with large data sets, since they may not properly scale when
data size grows.
Before using the models, they must be validated with a different set
of data than that used for estimation or training. The validation pro-
vides an estimation on the robustness of the models and the quality of
the predictions, so that the risk related to an inaccurate prediction can
be taken into account.
The methods used in this module may be applied in stream (i.e., the
models are continuously being trained with new data), scheduled (i.e.,
the models are trained periodically), or on demand (i.e, the user
manually requests to train again the models). Choosing one or other
approach depends mainly on the computational resources available for
this module.
The main output of the predictive techniques are the trained
models, whose application can guide the operative and the strategy of
the organization. They are made available to the rest of the organiza-
tion by means of the results publishing module.
6.1.6. Results publishing module
The purpose of this module is to provide the organization with a
decision-making tool. To do so, it makes the results of the analysis
conducted in the data analytics module available to the organization,
which includes the people that make decisions, but also other in-
formation systems that could benefit from the data analysis. For this
reason, this module should offer the results in different formats,
adapted to the different consumers of information in the organization.
The publication of results for decision-makers should be done in the
form of reports, including tables, graphics and other visual elements
that help understand the social and economic behavior behind the data.
The main objective of these reports is to support decisions at strategic
or tactical levels.
Making the analysis results available to other information systems
in the organization contributes to support the decision-making at op-
erational level. There is a wide variety of options to do so. For instance,
a trained model can be stored in a database or in any other storage for
being applied by different business units. The model could also be of-
fered as a service (under the SaaS paradigm) so that when a new event
occurs, the service offers the prediction as a result of applying the
model. This way, the trained models can be successfully applied to
some operational actions such as purchases and financial resources
management.
6.2. Governance layer
This layer is horizontal to the rest of the system and applies the
organization policies and regulations to the whole data lifecycle: from
the data ingestion to the disposal. It is composed of five modules, four
of them related to the data lifecycle, plus one for auditing purposes.
12

• Ingestion module: It deals with the management of the sources, in-
cluding the licenses and allowed uses, credentials for accessing
them, internal user permissions, completeness of metadata, and so
on.
• Processing module: It manages the privacy and anonymization po-
licies, controls processing for ethical principles, keeps track of the
transformations, as well as of the permissions for accessing the data
and computing resources.
• Results module: It is concerned with the traceability of the results
(from the sources to the final report), the permissions for accessing
the reports and results, along with the privacy aspects that may
affect the reports.
• Archival and disposal module: It implements the policy for archiving
and disposing the information related to data sources, processing
procedures and generated reports.
• Auditing module: It inspects that the implementation of the archi-
tecture is consistent with the current regulations, as well as with the
security and privacy policies. It may also include checking the
overall performance of the architecture in order to ensure that the
system has an acceptable response time.
6.3. Persistence layer
The persistence layer supports the other layers by managing all is-
sues related to the storage needs. Its main function is associated with
the storage of the data used as input in the data analysis layer, including
the schema for describing the relations among sources and other me-
tadata. Not only the data itself is covered, but also the storage of the
procedures followed in the different modules to access and transform
the data.
Furthermore, this layer serves the data analytics and results pub-
lishing modules by providing storage for the results. This includes
storing the models and providing them with the inputs required for
computing new predictions or estimations as part of the publication of
the results.
It is in the persistence layer where the storage infrastructure is
controlled and managed. This layer will typically use distributed sto-
rage systems, combining local storage with cloud solutions that allow
elastic storage and large volume data. The decision on whether to use
local or cloud storage mainly depends on where (on- or off-premises)
the modules intensive in computing power (e.g., data analytics) are
implemented.
7. Conclusions
In the Digital Era, most economic and social behaviors leaves be-
hind a huge digital footprint, which is incipiently being used with
nowcasting and forecasting purposes. Despite the enormous potential of
these data, integrating and analyzing the wide variety of heterogeneous
sources cannot be tackled with the traditional methods used in eco-
nomics and social sciences. To succeed in this purpose, it is mandatory
to carefully plan and implement the whole process of data extraction,
transformation and analysis. This is the point in which the Big Data and
data lifecycle paradigms arise as helpful perspectives on how to deal
with this process.
This paper has proposed a novel Big Data architecture that accounts
for the particularities of the economic and social behavior analyses in
the Digital Era. The first particularity is related to the variety of sources
that could provide information about economic and social topics. Our
first contribution addresses this issue by reviewing the multiple data
sources and proposing a taxonomy to classify them according to the
purpose of the agent generating the data.
Following the Big Data paradigm, this wide variety of hetero-
geneous sources requires specific methods for processing them. The
second contribution of the paper addresses this issue by reviewing those
methods not so commonly used in the social sciences, and classifying
them according to the phase of the data analysis they operate.
In order to frame the data analysis in an organizational perspective
and allow its management in a robust and flexible architecture, the data
lifecycle approach has been taken. Different perspectives on this ap-
proach have been reviewed and synthesized to establish and define all
the involved phases and processes.
Finally, the main contribution of the paper is the proposal of a Big
Data architecture adapted to the particularities of the economic and
social analyses, and grounded on the data lifecycle approach for the
management of data in the organization. At the same time, the proposal
aimed to be general enough to be implemented with different tech-
nologies, computing paradigms and analytical software depending on
the requirements and purposes of each particular case. By im-
plementing this architecture, an organization will be able to make the
most of all social and economic sources of information to which it has
access. Not only the organization of sources is advantageous, but also
their integration and connection to Big Data analytics tools able to run
the models for nowcasting and forecasting socio-economic variables.
The wide variety of data sources and techniques considered in the ar-
chitecture results in potentially more accurate and granular predictions.
Governments and official statistics institutions may also benefit
from the implementation of an information system with the proposed
architecture. Integrating the multiple sources to which they have access
may result in improved predictions about key economic indicators and
planning economic policies accordingly.
Although the proposed architecture is general enough to be im-
plemented with any technology, its adoption is not without obstacles.
To mention some of them, the integration of the architecture in the
existing organizational information systems is a critical process to en-
sure the smooth generation of forecasts and nowcasts. The im-
plementation of the modules in a proper cloud computing environment
so that the system can scale easily is also crucial. As future work, we
plan to implement the proposed Big Data architecture, in order to
generate and publish real-time nowcasts and forecasts of some socio-
economic variables using Internet data.
Acknowledgments
This work has been partially supported by the Spanish Ministry of
Economy and Competitiveness under Grant TIN2013-43913-R; and by
the Spanish Ministry of Education under Grant FPU14/02386.
References
Alis, C.M., Letchford, A., Moat, H.S., Preis, T., 2015. Estimating tourism statistics with
Wikipedia page views. In: WebSci-15 Proceedings of the ACM Web Science
Conference, Oxford, United Kingdom, http://dx.doi.org/10.1145/2786451.2786925.
Arenas-Márquez, F., Martínez-Torres, M., Toral, S., 2014. Electronic word-of-mouth
communities from the perspective of social network analysis. Tech. Anal. Strat.
Manag. 26 (8), 927–942. http://dx.doi.org/10.1080/09537325.2014.923565.
Armentano, M.G., Godoy, D., Campo, M., Amandi, A., 2014. NLP-based faceted search:
experience in the development of a science and technology search engine. Expert
Syst. Appl. 41 (6), 2886–2896. http://dx.doi.org/10.1016/j.eswa.2013.10.023.
Arora, S.K., Li, Y., Youtie, J., Shapira, P., 2016. Using the Wayback Machine to mine
websites in the social sciences: a methodological resource. J. Assoc. Inf. Sci. Technol.
67 (8), 1904–1915. http://dx.doi.org/10.1002/asi.23503.
Arora, S.K., Youtie, J., Shapira, P., Gao, L., Ma, T., 2013. Entry strategies in an emerging
technology: a pilot web-based study of graphene firms. Scientometrics 95 (3),
1189–1207. http://dx.doi.org/10.1007/s11192-013-0950-7.
Arrigo, E., Liberati, C., Mariani, P., 2016. A multivariate approach to Facebook data for
marketing communication. In: Proceedings of the 1st International Conference on
Advanced Research Methods and Analytics (CARMA 2016). UPV Press, Universitat
Politécnica de Valéncia, Valencia, Spain. http://dx.doi.org/10.4995/CARMA2016.
2016.2974.
Artola, C., Pinto, F., de Pedraza García, P., 2015. Can internet searches forecast tourism
inflows? Int. J. Manpow. 36 (1), 103–116. http://dx.doi.org/10.1108/IJM-12-2014-
0259.
Askitas, N., Zimmermann, K.F., 2009. Google econometrics and unemployment fore-
casting. Appl. Econ. Q. 55 (2), 107–120. http://dx.doi.org/10.3790/aeq.55.2.107.
Askitas, N., Zimmermann, K.F., 2013. Nowcasting business cycles using toll data. J.
Forecast. 32 (4), 299–306. http://dx.doi.org/10.1002/for.1262.
Askitas, N., Zimmermann, K.F., 2015. The internet as a data source for advancement in
13

social sciences. Int. J. Manpow. 36 (1), 2–12. http://dx.doi.org/10.1108/IJM-02-
2015-0029.
Assunção, M.D., Calheiros, R.N., Bianchi, S., Netto, M.A., Buyya, R., 2015. Big Data
computing and clouds: trends and future directions. J. Parallel Distrib. Comput. 79-
80, 3–15. http://dx.doi.org/10.1016/j.jpdc.2014.08.003.
Bahrami, M., Singhal, M., 2014. The Role of Cloud Computing Architecture in Big Data.
vol. 8. Springer International Publishing, Cham, pp. 275–295. http://dx.doi.org/10.
1007/978-3-319-08254-7_13.
Bangwayo-Skeete, P.F., Skeete, R.W., 2015. Can Google data improve the forecasting
performance of tourist arrivals? Mixed-data sampling approach. Tour. Manag. 46,
454–464. http://dx.doi.org/10.1016/j.tourman.2014.07.014.
Bello-Orgaz, G., Jung, J.J., Camacho, D., 2016. Social big data: recent achievements and
new challenges. Inf. Fusion 28, 45–59. http://dx.doi.org/10.1016/j.inffus.2015.08.
005.
Berman, F., Fox, G., Hey, A.J., 2003. Grid Computing: Making the Global Infrastructure a
Reality. Communications Networking & Distributed Systems John Wiley and Sons.
Blazquez, D., Domenech, J., 2017. Web data mining for monitoring business export or-
ientation. Technol. Econ. Dev. Econ. Online, 1–23. http://dx.doi.org/10.3846/
20294913.2016.1213193.
Blei, D.M., Ng, A.Y., Jordan, M.I., 2003. Latent Dirichlet allocation. J. Mach. Learn. Res.
3, 993–1022.
Bollen, J., Mao, H., Zeng, X., 2011. Twitter mood predicts the stock market. J. Comput.
Sci. 2 (1), 1–8. http://dx.doi.org/10.1016/j.jocs.2010.12.007.
Cameron, M.P., Barrett, P., Stewardson, B., 2016. Can social media predict election re-
sults? Evidence from New Zealand. J. Polit. Mark. 15 (4), 416–432. http://dx.doi.
org/10.1080/15377857.2014.959690.
Ceron, A., Curini, L., Iacus, S.M., Porro, G., 2014. Every tweet counts? How sentiment
analysis of social media can improve our knowledge of citizens' political preferences
with an application to Italy and France. New Media Soc. 16 (2), 340–358. http://dx.
doi.org/10.1177/1461444813480466.
Ceron, A., Negri, F., 2016. The “social side” of public policy: monitoring online public
opinion and its mobilization during the policy cycle. Policy Internet 8 (2), 131–147.
http://dx.doi.org/10.1002/poi3.117.
Chapman, P., Clinton, J., Kerber, R., Khabaza, T., Reinartz, T., Shearer, C., Wirth, R.,
2000. CRISP-DM 1.0 - Step-by-Step Data Mining Guide. https://www.the-modeling-
agency.com/crisp-dm.pdf (accessed 1st June, 2017).
Chen, M., Mao, S., Liu, Y., 2014. Big Data: a survey. Mob. Netw. Appl. 19 (2), 171–209.
http://dx.doi.org/10.1007/s11036-013-0489-0.
Chittaranjan, G., Blom, J., Gatica-Perez, D., 2013. Mining large-scale smartphone data for
personality studies. Pers. Ubiquit. Comput. 17 (3), 433–450. http://dx.doi.org/10.
1007/s00779-011-0490-1.
Choi, H., Varian, H., 2009a. Predicting Initial Claims for Unemployment Benefits. http://
research.google.com/archive/papers/initialclaimsUS.pdf (accessed 10th October,
2016).
Choi, H., Varian, H., 2009b. Predicting the Present with Google Trends. http://static.
googleusercontent.com/external_content/untrusted_dlcp/www.google.com/en//
googleblogs/pdfs/google_predicting_the_present.pdf (accessed 10th October, 2016).
Chong, A.Y.L., Ch’ng, E., Liu, M.J., Li, B., 2015. Predicting consumer product demands via
Big Data: the roles of online promotional marketing and online reviews. Int. J. Prod.
Res. Online, 1–15. http://dx.doi.org/10.1080/00207543.2015.1066519.
Chou, J.-S., Ngo, N.-T., 2016. Smart grid data analytics framework for increasing energy
savings in residential buildings. Autom. Constr. 72 (3), 247–257. http://dx.doi.org/
10.1016/j.autcon.2016.01.002.
Chowdhury, G.G., 2005. Natural language processing. Annu. Rev. Inf. Sci. Technol. 37
(1), 51–89. http://dx.doi.org/10.1002/aris.1440370103.
Committee on Earth Observation Satellites — Working Group on Information Systems and
Services, 2012. CEOS Data Life Cycle Models and Concepts. https://my.usgs.gov/
confluence/download/attachments/82935852/Data%20Lifecycle%20Models
%20and%20Concepts% (accessed 27th September, 2016).
Congdon, P., 2007. Bayesian Statistical Modelling. Wiley Series in Probability and
Statistics, 2nd. John Wiley & Sons.
Cortes, C., Vapnik, V., 1995. Support-vector networks. Mach. Learn. 20 (3), 273–297.
http://dx.doi.org/10.1007/BF00994018.
Corti, L., Van den Eynden, V., Bishop, L., Woollard, M., 2014. Managing and sharing
research data: a guide to good practice, 1st. Sage Publications.
Cox, M., Ellsworth, D., 1997. Managing Big Data for scientific visualization. ACM
Siggraph, MRJ/NASA Ames Res. Cent. 5, 1–17.
David, E., Zhitomirsky-Geffet, M., Koppel, M., Uzan, H., 2016. Utilizing Facebook pages
of the political parties to automatically predict the political orientation of Facebook
users. Online Inf. Rev. 40 (5), 610–623. http://dx.doi.org/10.1108/OIR-09-2015-
0308.
Alliance, D.D.I., 2008. DDI Lifecycle 3.0. http://www.ddialliance.org/ (accessed 29th
September, 2016).
Deville, P., Linard, C., Martin, S., Gilbert, M., Stevens, F.R., Gaughan, A.E., Blondel, V.D.,
Tatem, A.J., 2014. Dynamic population mapping using mobile phone data. Proc. Natl.
Acad. Sci. 111 (45), 15888–15893. http://dx.doi.org/10.1073/pnas.1408439111.
Dey, M.M., Rabbani, A.G., Singh, K., Engle, C.R., 2014. Determinants of retail price and
sales volume of catfish products in the United States: an application of retail scanner
data. Aquac. Econ. Manag. 18 (2), 120–148. http://dx.doi.org/10.1080/13657305.
2014.903312.
Dobra, A., Williams, N.E., Eagle, N., 2015. Spatiotemporal detection of unusual human
population behavior using mobile phone data. PLOS ONE 10 (3), 1–20. http://dx.doi.
org/10.1371/journal.pone.0120449.s001.
Domenech, J., de la Ossa, B., Pont, A., Gil, J.A., Martinez, M., Rubio, A., 2012. An in-
telligent system for retrieving economic information from corporate websites. In:
IEEE/WIC/ACM International Joint Conferences on Web Intelligence (WI) and
Intelligent Agent Technologies (IAT), Macau, China, pp. 573–578. http://dx.doi.org/
10.1109/WI-IAT.2012.92.
Edelman, B., 2012. Using Internet data for economic research. J. Econ. Perspect. 26 (2),
189–206. http://dx.doi.org/10.1257/jep.26.2.189.
Einav, L., Levin, J., 2014. The data revolution and economic analysis. Innov. Policy Econ.
14 (1), 1–24. http://dx.doi.org/10.1086/674019.
Evangelopoulos, N., Zhang, X., Prybutok, V.R., 2012. Latent semantic analysis: five
methodological recommendations. Eur. J. Inf. Syst. 21 (1), 70–86. http://dx.doi.org/
10.1057/ejis.2010.61.
Fan, J., Han, F., Liu, H., 2014. Challenges of big data analysis. Nat. Sci. Rev. 1 (2),
293–314. http://dx.doi.org/10.1093/nsr/nwt032.
Fawcett, T., 2006. An introduction to ROC analysis. Pattern Recogn. Lett. 27 (8),
861–874. http://dx.doi.org/10.1016/j.patrec.2005.10.010.
Fayyad, U., Piatetsky-Shapiro, G., Smyth, P., 1996. The KDD process for extracting useful
knowledge from volumes of data. Commun. ACM 39 (11), 27–34.
Fondeur, Y., Karamé, F., 2013. Can Google data help predict French youth unemploy-
ment? Econ. Model. 30, 117–125. http://dx.doi.org/10.1016/j.econmod.2012.07.
017.
Friedman, J., Hastie, T., Tibshirani, R., 2010. Regularization paths for Generalized Linear
Models via coordinate descent. J. Stat. Softw. 33 (1), 1–22.
Frota Neto, J.Q., Bloemhof, J., Corbett, C., 2016. Market prices of remanufactured, used
and new items: evidence from eBay. Int. J. Prod. Econ. 171 (3), 371–380. http://dx.
doi.org/10.1016/j.ijpe.2015.02.006.
Gaikar, D.D., Marakarkandy, B., Dasgupta, C., 2015. Using Twitter data to predict the
performance of Bollywood movies. Ind. Manag. Data Syst. 115 (9), 1604–1621.
http://dx.doi.org/10.1108/IMDS-04-2015-0145.
Gandomi, A., Haider, M., 2015. Beyond the hype: big data concepts, methods, and ana-
lytics. Int. J. Inf. Manag. 35 (2), 137–144. http://dx.doi.org/10.1016/j.ijinfomgt.
2014.10.007.
Gayo-Avello, D., 2012, Nov, Nov. No, you cannot predict elections with twitter. IEEE
Internet Comput. 16 (6), 91–94. http://dx.doi.org/10.1109/MIC.2012.137.
Gayo-Avello, D., 2013. A meta-analysis of state-of-the-art electoral prediction from
Twitter data. Soc. Sci. Comput. Rev. 31 (6), 649–679. http://dx.doi.org/10.1177/
0894439313493979.
Gök, A., Waterworth, A., Shapira, P., 2015. Use of web mining in studying innovation.
Scientometrics 102 (1), 653–671. http://dx.doi.org/10.1007/s11192-014-1434-0.
Graells-Garrido, E., Peredo, O., García, J., 2016. Sensing urban patterns with antenna
mappings: the case of Santiago, Chile. Sensors 16 (7), 1098–1123. http://dx.doi.org/
10.3390/s16071098.
Hand, C., Judge, G., 2012. Searching for the picture: forecasting UK cinema admissions
using Google Trends data. Appl. Econ. Lett. 19 (11), 1051–1055. http://dx.doi.org/
10.1080/13504851.2011.613744.
Harsanyi, J.C., 1978. Bayesian decision theory and utilitarian ethics. Am. Econ. Rev. 68
(2), 223–228.
Hashem, I.A.T., Yaqoob, I., Anuar, N.B., Mokhtar, S., Gani, A., Ullah Khan, S., 2015. The
rise of “big data” on cloud computing: review and open research issues. Inf. Syst. 47,
98–115. http://dx.doi.org/10.1016/j.is.2014.07.006.
Hastie, T., Rosset, S., Tibshirani, R., Zhu, J., 2004. The entire regularization path for the
support vector machine. J. Mach. Learn. Res. 5, 1391–1415.
Hastie, T., Tibshirani, R., Friedman, J., 2013. The Elements of Statistical Learning: Data
Mining, Inference and Prediction. Springer Series in Statistics, 3rd. Springer.
He, H., Garcia, E., 2009. Learning from imbalanced data. IEEE Trans. Knowl. Data Eng. 21
(9), 1263–1284. http://dx.doi.org/10.1109/TKDE.2008.239.
Hu, N., Bose, I., Koh, N.S., Liu, L., 2012. Manipulation of online reviews: an analysis of
ratings, readability, and sentiments. Decis. Support. Syst. 52 (3), 674–684. http://dx.
doi.org/10.1016/j.dss.2011.11.002.
IBM, 2016. Big Data and Analytics. http://www-01.ibm.com/software/data/bigdata/
what-is-big-data.html (accessed 21st December, 2016).
Jagadish, H., Gehrke, J., Labrinidis, A., Papakonstantinou, Y., Patel, J.M., Ramakrishnan,
R., Shahabi, C., 2014. Big data and its technical challenges. Commun. ACM 57 (7),
86–94. http://dx.doi.org/10.1145/2611567.
Jin, X., Wah, B.W., Cheng, X., Wang, Y., 2015. Significance and challenges of big data
research. Big Data Res. 2 (2), 59–64. http://dx.doi.org/10.1016/j.bdr.2015.01.006.
Khadivi, P., Ramakrishnan, N., 2016. Wikipedia in the tourism industry: forecasting de-
mand and modeling usage behavior. In: Thirtieth AAAI Conference on Artificial
Intelligence. February 12–17, 2016, Phoenix, Arizona, pp. 4016–4021.
Kim, M., Park, H.W., 2012. Measuring Twitter-based political participation and delib-
eration in the South Korean context by using social network and Triple Helix in-
dicators. Scientometrics 90 (1), 121–140. http://dx.doi.org/10.1007/s11192-011-
0508-5.
Kim, T., Hong, J., Kang, P., 2015. Box office forecasting using machine learning algo-
rithms based on SNS data. Int. J. Forecast. 31 (2), 364–390. http://dx.doi.org/10.
1016/j.ijforecast.2014.05.006.
Kitchin, R., 2014. The real-time city? Big data and smart urbanism. GeoJournal 79 (1),
1–14. http://dx.doi.org/10.1007/s10708-013-9516-8.
Krishnan, N.C., Cook, D.J., 2014. Activity recognition on streaming sensor data. Pervasive
Mob. Comput. 10, 138–154. http://dx.doi.org/10.1016/j.pmcj.2012.07.003.
Laney, D., 2001. 3D Data Management: Controlling Data Volume, Velocity, and Variety.
Application Delivery Strategies. pp. 949. http://blogs.gartner.com/doug-laney/files/
2012/01/ad949-3D-Data-Management-Controlling-Data-Volume-Velocity-and-
Variety.pdf (accessed 21st December, 2016).
Laurila, J.K., Gatica-Perez, D., Aad, I., Blom, J., Bornet, O., Do, T.M.T., Dousse, O., Eberle,
J., Miettinen, M., 2013. From big smartphone data to worldwide research: the mobile
data challenge. Pervasive Mob. Comput. 9, 752–771. http://dx.doi.org/10.1016/j.
pmcj.2013.07.014.
LeCun, Y., Bengio, Y., Hinton, G., 2015. Deep learning. Nature 521 (7553), 436–444.
14

http://dx.doi.org/10.1038/nature14539.
Ley, E., Steel, M.F., 2012. Mixtures of g-priors for Bayesian model averaging with eco-
nomic applications. J. Econ. 171 (2), 251–266. http://dx.doi.org/10.1016/j.jeconom.
2012.06.009.
Li, G., Law, R., Vu, H.Q., Rong, J., Zhao, X.R., 2015. Identifying emerging hotel pre-
ferences using Emerging Pattern Mining technique. Tour. Manag. 46, 311–321.
http://dx.doi.org/10.1016/j.tourman.2014.06.015.
Li, Y., Arora, S., Youtie, J., Shapira, P., 2016. Using web mining to explore Triple Helix
influences on growth in small and mid-size firms. Technovation Online, 1–12. http://
dx.doi.org/10.1016/j.technovation.2016.01.002.
Liu, B., 2012. Sentiment analysis and opinion mining. Synth. Lect. Hum. Lang. Technol. 5
(1), 1–167. http://dx.doi.org/10.2200/S00416ED1V01Y201204HLT016.
Liu, Y., Huang, X., An, A., Yu, X., 2007. ARSA: sentiment-aware model for predicting sales
performance using blogs. In: Proceedings of the 30th Annual International ACM
SIGIR Conference on Research and Development in Information Retrieval. ACM, pp.
607–614.
Ludwig, O., Nunes, U., Araujo, R., 2014. Eigenvalue decay: a new method for neural
network regularization. Neurocomputing 124, 33–42. http://dx.doi.org/10.1016/j.
neucom.2013.08.005.
Lynch, C., 2008. Big data: how do your data grow? Nature 455, 28–29. http://dx.doi.org/
10.1038/455028a.
Malbon, J., 2013. Taking fake online consumer reviews seriously. J. Consum. Policy 36
(2), 139–157. http://dx.doi.org/10.1007/s10603-012-9216-7.
Mavragani, A., Tsagarakis, K.P., 2016. YES or NO: predicting the 2015 GReferendum
results using Google Trends. Technol. Forecast. Soc. Chang. 109, 1–5. http://dx.doi.
org/10.1016/j.techfore.2016.04.028.
McLaren, N., Shanbhogue, R., 2011. Using internet search data as economic indicators.
Bank Engl. Q. Bull. 2011 Q2, 134–140.
Menardi, G., Torelli, N., 2014. Training and assessing classification rules with imbalanced
data. Data Min. Knowl. Disc. 28, 92–122. http://dx.doi.org/10.1007/s10618-012-
0295-5.
Moat, H.S., Curme, C., Stanley, H.E., Preis, T., 2014. Anticipating Stock Market
Movements with Google and Wikipedia. NATO Science for Peace and Security Series
C: Environmental Security Springer Science, pp. 47–59. http://dx.doi.org/10.1007/
978-94-017-8704-8_4.
Montoliu, R., Blom, J., Gatica-Perez, D., 2013. Discovering places of interest in everyday
life from smartphone data. Multimedia Tools Appl. 62, 179–207. http://dx.doi.org/
10.1007/s11042-011-0982-z.
Moro, S., Cortez, P., Rita, P., 2015. Business intelligence in banking: a literature analysis
from 2002 to 2013 using text mining and latent Dirichlet allocation. Expert Syst.
Appl. 42, 1314–1324. http://dx.doi.org/10.1016/j.eswa.2014.09.024.
Najafabadi, M.M., Villanustre, F., Khoshgoftaar, T.M., Seliya, N., Wald, R., Muharemagic,
E., 2015. Deep learning applications and challenges in big data analytics. J. Big Data
2 (1), 1–21. http://dx.doi.org/10.1186/s40537-014-0007-7.
Pääkkönen, P., Pakkala, D., 2015. Reference architecture and classification of technolo-
gies, products and services for big data systems. Big Data Res. 2 (4), 166–186. http://
dx.doi.org/10.1016/j.bdr.2015.01.001.
Pan, B., Yang, Y., 2016. Monitoring and Forecasting Tourist Activities with Big Data.
Apple Academic Press, pp. 43–62 chap. 3.
Pandya, S.S., Venkatesan, R., 2016. French roast: consumer response to international
conflict — evidence from supermarket scanner data. Rev. Econ. Stat. 98 (1), 42–56.
http://dx.doi.org/10.1162/REST_a_00526.
Pesenson, M.Z., Pesenson, I.Z., McCollum, B., 2010. The data big bang and the expanding
digital universe: high-dimensional, complex and massive data sets in an inflationary
epoch. Adv. Astron. 2010, 1–16. http://dx.doi.org/10.1155/2010/350891.
Preis, T., Moat, H.S., Stanley, H.E., 2013. Quantifying trading behavior in financial
markets using Google Trends. Sci Rep 3, 1–6. http://dx.doi.org/10.1038/srep01684.
Reed, D.A., Dongarra, J., 2015. Exascale computing and big data. Commun. ACM 58,
56–68. http://dx.doi.org/10.1145/2699414.
Rudolph, M., Ruiz, F., Mandt, S., Blei, D., 2016. Exponential family embeddings. In:
Advances in Neural Information Processing Systems, pp. 478–486.
Rüegg, J., Gries, C., Bond-Lamberty, B., Bowen, G.J., Felzer, B.S., McIntyre, N.E.,
Soranno, P.A., Vanderbilt, K.L., Weathers, K.C., 2014. Completing the data life cycle:
using information management in macrosystems ecology research. Front. Ecol.
Environ. 12, 24–30. http://dx.doi.org/10.1890/120375.
Russell, M.A., 2013. Mining the Social Web: Data Mining Facebook, Twitter, LinkedIn,
Google+, GitHub and More, 2nd. O’Reilly Media, pp. 448.
Saleiro, P., Amir, S., Silva, M., Soares, C., 2015. Popmine: tracking political opinion on the
web. In: 2015 IEEE International Conference on Computer and Information
Technology; Ubiquitous Computing and Communications; Dependable, Autonomic
and Secure Computing; Pervasive Intelligence and Computing (CIT/IUCC/DASC/
PICOM). IEEE, Liverpool, United Kingdom, pp. 1521–1526.
Schneider, M.J., Gupta, S., 2016. Forecasting sales of new and existing products using
consumer reviews: a random projections approach. Int. J. Forecast. 32, 243–256.
http://dx.doi.org/10.1016/j.ijforecast.2015.08.005.
Schoen, H., Gayo-Avello, Panagiotis Takis Metax, D., Takis Metaxas, P., Mustafaraj, E.,
Strohmaier, M., 2013. The power of prediction with social media. Internet Res. 23,
528–543. http://dx.doi.org/10.1108/IntR-06-2013-0115.
Simonet, A., Fedak, G., Ripeanu, M., 2015. Active data: a programming model to manage
data life cycle across heterogeneous systems and infrastructures. Futur. Gener.
Comput. Syst. 53, 25–42. http://dx.doi.org/10.1016/j.future.2015.05.015.
Suhara, Y., Xu, Y., Pentland, A., 2017. Deepmood: forecasting depressed mood based on
self-reported histories via recurrent neural networks. In: Proceedings of the 26th
International Conference on World Wide Web - WWW ’17. ACM Press, pp. 715–724.
Sun, Y., Kamel, M.S., Wong, A.K., Wang, Y., 2007. Cost-sensitive boosting for classifica-
tion of imbalanced data. Pattern Recogn. 40 (12), 3358–3378. http://dx.doi.org/10.
1016/j.patcog.2007.04.009.
Thelwall, M., 2007. Blog searching: the first general-purpose source of retrospective
public opinion in the social sciences? Online Inf. Rev. 31 (3), 277–289. http://dx.doi.
org/10.1108/14684520710764069.
Tibshirani, R., 1996. Regression shrinkage and selection via the Lasso. J. R. Stat. Soc. Ser.
B Methodol. 58, 267–288.
Tumasjan, A., Sprenger, T.O., Sandner, P.G., Welpe, I.M., 2011. Election forecasts with
Twitter: how 140 characters reflect the political landscape. Soc. Sci. Comput. Rev. 29
(4), 402–418. http://dx.doi.org/10.1177/0894439310386557.
Van Vlasselaer, V., Bravo, C., Caelen, O., Eliassi-Rad, T., Akoglu, L., Snoeck, M., Baesens,
B., 2015. APATE: a novel approach for automated credit card transaction fraud de-
tection using network-based extensions. Decis. Support. Syst. 75, 38–48. http://dx.
doi.org/10.1016/j.dss.2015.04.013.
Varian, H.R., 2014. Big Data: new tricks for econometrics. J. Econ. Perspect. 28, 3–28.
http://dx.doi.org/10.1257/jep.28.2.3.
Vatsalan, D., Christen, P., Verykios, V.S., 2013. A taxonomy of privacy-preserving record
linkage techniques. Inf. Syst. 38, 946–969. http://dx.doi.org/10.1016/j.is.2012.11.
005.
Vicente, M.R., López-Menéndez, A.J., Pérez, R., 2015. Forecasting unemployment with
internet search data: does it help to improve predictions when job destruction is
skyrocketing? Technol. Forecast. Soc. Chang. 92, 132–139. http://dx.doi.org/10.
1016/j.techfore.2014.12.005.
Vosen, S., Schmidt, T., 2011. Forecasting private consumption: survey-based indicators
vs. Google Trends. J. Forecast. 30 (6), 565–578. http://dx.doi.org/10.1002/for.1213.
Wang, Y., Kung, L., Byrd, T.A., 2016a. Big data analytics: understanding its capabilities
and potential benefits for healthcare organizations. Technol. Forecast. Soc. Chang.
Online, 1–11. http://dx.doi.org/10.1016/j.techfore.2015.12.019.
Wang, Y., Yuan, N.J., Sun, Y., Zhang, F., Xie, X., Li, Q., Chen, E., 2016b. A contextual
collaborative approach for app usage forecasting. In: Proceedings of the 2016 ACM
International Joint Conference on Pervasive and Ubiquitous Computing — UbiComp
’16. ACM Press, pp. 1247–1258.
Wikimedia Foundation, 2017. Dashboards and Data Downloads for Wikimedia Projects.
https://analytics.wikimedia.org/ (accessed 7th July, 2017).
Williams, N.E., Thomas, T.A., Dunbar, M., Eagle, N., Dobra, A., 2015. Measures of human
mobility using mobile phone records enhanced with GIS data. PLOS ONE 10, 1–16.
http://dx.doi.org/10.1371/journal.pone.0133630.
Witten, I.H., Frank, E., Hall, M.A., Pal, C.J., 2016. Data Mining: Practical Machine
Learning Tools and Techniques, 4th. Morgan Kaufmann - Elsevier.
Wu, J., Pan, S., Zhu, X., Cai, Z., Zhang, P., Zhang, C., 2015. Self-adaptive attribute
weighting for Naive Bayes classification. Expert Syst. Appl. 42 (3), 1487–1502.
http://dx.doi.org/10.1016/j.eswa.2014.09.019.
Xiong, T., Wang, S., Mayers, A., Monga, E., 2013. Personal bankruptcy prediction by
mining credit card data. Expert Syst. Appl. 40, 665–676. http://dx.doi.org/10.1016/
j.eswa.2012.07.072.
Zhang, Y., Ren, S., Liu, Y., Si, S., 2017. A big data analytics architecture for cleaner
manufacturing and maintenance processes of complex products. J. Clean. Prod. 142
(2), 626–641. http://dx.doi.org/10.1016/j.jclepro.2016.07.123.
Zou, H., Hastie, T., 2005. Regularization and variable selection via the elastic net. J. R.
Stat. Soc. Ser. B 67, 301–320. http://dx.doi.org/10.1111/j.1467-9868.2005.00503.x.
Desamparados Blazquez received a B.S. in Business Administration and Management
and a M.S. in Data Analytics Engineering from the Universitat Politècnica de València
(Spain). She is currently a Ph.D. student and a predoctoral research fellow under the
Programme for the Training of University Lecturers (FPU) from the Spanish Ministry of
Education. She develops her work at the Department of Economics and Social Sciences of
the Universitat Politècnica de València. Her research interests include web economic
indicators and internet economics.
Josep Domenech received a B.S., M.S. and Ph.D. in Computer Science from the
Universitat Politècnica de València, and a B.S. and M.S. in Business Administration and
Economics from the Universitat de València. Since 2009, he is an associate professor at
the Department of Economics and Social Sciences of the Universitat Politècnica de
València. He is currently leading a research project on planning and using cloud com-
puting platforms efficiently, funded by the Ministry of Economy and Competitiveness of
Spain. His research interests are focused on multidisciplinary approaches to internet
systems and digital economics, including web economic indicators, internet economics
and web performance characterization.
15

Big data sources and methods for social and economic analyses

More Related Content

What's hot (19)

Similar to Big data sources and methods for social and economic analyses (20)

More from Americo Arizaca Avalos (13)

Recently uploaded (20)

Big data sources and methods for social and economic analyses