CHAPTER

4
Introduction to Social Media
Analytics

Dr. Gohar F. Khan

School of Management and Marketing, University of Waikato

“Analyzing is believing” —Gohar F. Khan

1. Introduction
Social Media Analytics is the art and science of extracting valuable hidden
insights from vast amounts of semi-structured and unstructured social media data
to enable informed and insightful decision-making. It is considered a science,
as it involves systematically identifying, extracting, and analyzing social media
data (such as tweets, shares, likes, and hyperlinks) using sophisticated tools and
techniques. It is also an art, interpreting and aligning the insights gained with
business goals and objectives. To get value from analytics, one should master both
its art and science.
The science part of social media analytics requires skilled data analysts,
sophisticated tools and technologies, and reliable data. Getting the science right,
however, is not enough. To effectively consider the results and put them into
action, the business must master the other half of analytics, that is, the art of
interpreting and aligning analytics with business objectives and goals. Interpreting
analytics results, for example, requires representing the data in meaningful ways,
having domain-specific knowledge, and training. Analytics should be strategically
aligned to support existing business goals. Without a well-crafted and aligned
social media strategy, the business will struggle to get the desired outcomes
from analytics.

2. Emergence of Social Media Analytics

Social media analytics is a relatively new but emerging field. Based on Google’s
trends data (Figure 1), the term social media analytics seems to have appeared
66 Social Media Analytics in Predicting Consumer Behavior

Figure 1. Interest in social media analytics over time

over the Internet horizon in July 2006, and interest in it (that is people searching
for it) has steadily increased since then.
Google trends also show that majority of the interest in social media analytics
is coming from India, Canada, the United States and the United Kingdom, and
users who are interested in social media analytics also searched for a variety of
topics including Google analytics, social media marketing, social media tools,
marketing analytics, Facebook analytics, Twitter analytics, and social media
management. A full list of social media related terms is shown in Table 1. The
scoring shown in Table 1 is on a relative scale where a value of 100 is the most

Table 1. Social media analytics related terms

Terms Searched Score Terms Searched Score

Google analytics 100 Social media management 10
Social media marketing 40 Hootsuite analytics 10
Social media tools 40 SEO 10
Social analytics tools 35 Hootsuite 10
Marketing analytics 35 Instagram analytics 10
Facebook analytics 30 Big data analytics 10
Twitter analytics 30 Social media metrics 10
Data analytics 25 LinkedIn 5
Web analytics 20 Big data 5
Social media analysis 15 LinkedIn analytics 5
Social media tracking 10 WordPress analytics 5
Social media monitoring 10 Social media strategy 5
Social media management 10 Social media dashboard 5
Introduction to Social Media Analytics 67

commonly searched query, 50 is a query searched half as often, and a value of 0 is

a query searched for less than 1% as often as the most popular query.
As social media is becoming more mainstream, and people are using it to
express their feelings, and interests, share content, and, collaborate, therefore
the social media analytics field is also gaining prominence in both the research
and business communities. Businesses need to tap into the vast amounts of data
produced by social media users to increase brand loyalty, generate leads, drive
traffic, make forecasts, and ultimately make the right decisions. Social media
data and users are of significant value to businesses. A study, for example, found
that the average value of a Facebook fan was $174.17 in major consumer areas
(Syncapse 2013). KINAXIS, a supply chain management company, for example,
used eighteen employee bloggers and generated over forty-two million leads
(Petersen 2012).

3. Purpose of Social Media Analytics

The main premise of social media analytics is to enable informed and
insightful decision making by leveraging social media data (Chen et al. 2012);
(Bekmamedova and Shanks 2014). Businesses use social media for a variety of
reasons which include the following.
• Connecting and engaging with current customers
• Finding and engaging with new customers
• Getting feedback on products/services
• Generating business leads
• Driving traffic to business channels (Facebook pages, corporate blogs,
company web, etc.)
• To measure brand loyalty
• Tracking products/services/and campaign impact
• Predictive business forecasting
• Business intelligence and market research
In addition, the following are some sample questions that can be answered
with social media analytics.
• What are the customers who are using social media saying about our brand or
a new product launch?
• Which content posted on social media is resonating more with clients or
customers?
• How can I harness social media data (e.g., tweets and Facebook comments)
to improve our product/services?
• Is the social media conversation about our company, product, or service
positive, negative, or neutral?
• How can I leverage social media to promote brand awareness?
• Who are our influential social media followers, fans, and friends?
68 Social Media Analytics in Predicting Consumer Behavior

• Who are our influential social media nodes (e.g., people and organizations)
and their position in the network?
• Which social media platforms are driving the most traffic to our corporate
website?
• Where is the geographical location of our social media customers?
• Which keywords and terms are trending over social media?
• How active is social media in our business and how many people are
connected with us?
• Which websites are linked to our corporate website?
• How are my competitors doing on social media?
When it comes to social media data and using it to generate business value,
the statement at the beginning of the chapter can be no more than correct. In the
context of social media, seeing is no more than believing, rather analyzing it is.
In other words, business (and social and political) decisions should be based on
digging deep into the social media data rather than just by believing what we see
over social media. On the social web, each second tons of data are generated,
which may carry potential business insights; however, not all the social data is
gold. A vast amount of social data is either fake or useless. To separate good data
from bad, social media analytics coupled with human judgement is the answer.

4. Social Media vs. Conventional Business

Analytics
While the premise of both social media and traditional business analytics is to
produce actionable business, they do however slightly differ in scope and nature.
Table 2 below provides a comparison of social media analytics with traditional
business analytics. As an emerging field, it may not be appropriate to use the term
conventional for business analytics; we do so here only for comparison purposes.
The most visible difference between the two comes from the source, type,
and nature of the data which has been mined. Unlike the traditional business
analytics of structured and historical information, social media analytics involves
the collection, analysis, and interpretation of semistructured and unstructured
social media data to gain insight into contemporary issues while supporting
effective decision making (Bekmamedova and Shanks 2014). Social media data
is highly diverse, high volume, real-time, and stored in third-party databases in
a semi-structured and unstructured format. Structured business data, on the other
hand, is mostly stored in databases and spreadsheets in a machine-readable format
(e.g., rows and columns). Thus it can be easily searched, computed, and mined.
Unstructured and semi-structured social media data is not machine readable and
can take a variety of forms, such as the contents of this book, Facebook comments,
emails, tweets, hyperlinks, PowerPoint presentations, images, emoticons, videos,
etc. Thus, it is not analytics-friendly and needs a lot of cleaning and transformation.
Introduction to Social Media Analytics 69

Table 2. Social media analytics vs. business analytics

Social Media Analytics Business Analytics

Semi-structured and unstructured data Structured data
Data is not analytical friendly Data is analytical friendly
Real-time data Mostly historical data
Public data Private data
Stored in third-party databases Stored in business-owned databases
Boundary-less data (i.e., Boundary within the Bound within the business intranet
Internet)
Data is high volume Data is medium to high volume
Highly diverse data Uniform data
Data is widely shared over the Internet Data is only shared within organizations
More sharing creates greater value/impact Less sharing creates more value
No business control over data Tightly controlled by business
Socialized data Bureaucratic data
Data is informal in nature Data is formal in nature

Another visible difference comes from the way the information (i.e., text,
photographs, videos, audio, etc.) is created and consumed. Social media data
originates from the public Internet and is socialized by nature. Socialized data is
provided for the collective good. It is created and consumed using various social
media platforms and social technologies to maintain social and professional
ties (e.g., Facebook, LinkedIn, etc.), and to facilitate knowledge sharing and
management (Wikipedia, blogs, etc.). Socialized data creates awareness (i.e.,
Twitter), or is used to exchange information in the form of text, audio, video,
documents, and graphics, to name a few (Khan 2013). Social media data is social,
informal, and not bound (i.e., the Internet is a boundary), unlike conventional
analytics data, which is bureaucratic, formal in nature, controlled by organizations,
and bound or trapped within the organizational network or intranet. More
importantly, the value or impact of socialized data is determined by the extent
to which it is shared with other social entities (e.g., people or organizations): the
more it is shared (i.e., socialized) the greater its value.
For example, the value/effect of information can be measured by attracting
more followers (e.g., on Twitter or Facebook). Another measure is page views or
clicks, or regarding socio-political impact (e.g., information disseminated using
social media to organize political or social movements may have more effect
regarding organizing the events). However, the majority of the conventional
business data is confined within organizational databases, limitedly shared, and
can serve as a source of competitive advantage.
70 Social Media Analytics in Predicting Consumer Behavior

Eight Layers of Social Media Analytics

Social media at a minimum has eight layers of data (Figure 2). Each layer
carries potentially valuable information and insights that can be harvested for
business intelligence purposes. Out of the eight layers, some are visible or easily
identifiable (e.g., text and actions), and others are invisible (e.g., social media
and hyperlink networks). Most of these layers can be used to develop measurable
KPIs. The following are eight layers of social media analytics (Khan 2018) (for a
detailed discussion on these layers, please refer to Khan’s 2018 book, “Creating
Value With Social Media Analytics: Managing, Aligning, and Mining Social
Media Text, Networks, Actions, Location, Apps, Hyperlinks, Multimedia, Search
Engines Data”). Here we briefly outline these layers.
1. Text
2. Networks
3. Actions
4. Hyperlinks
5. Mobile
6. Location
7. Multimedia
8. Search engines

Layer One: Text

Social media text analytics deals with the extraction and analysis of business
insights from textual elements of social media content, such as comments,
tweets, blog posts, and, Facebook status updates. Text analytics is mostly used to
understand social media users’ sentiments or identify emerging themes and topics.

Figure 2. Eight layers of social media analytics (Source: Khan 2018)

Introduction to Social Media Analytics 71

Layer Two: Networks

Social media network analytics extract, analyze, and interpret personal and
professional social networks, for example, Facebook, Friendship Network, and
Twitter. Network analytics seeks to identify influential nodes (e.g., people and
organizations) and their position in the network.

Layer Three: Actions

Social media actions analytics deals with extracting, analyzing, and interpreting
the actions performed by social media users, including likes, dislikes, shares,
mentions and, endorsements. Actions analytics are mostly used to measure
popularity, influence, and prediction in social media.

Layer Four: Search Engines

Search engine analytics focuses on analyzing historical search data for gaining
valuable insight into a range of areas, including trends analysis, keyword
monitoring, search result and advertisement history, and advertisement
spending statistics.

Layer Five: Location

Location analytics, also known as spatial analysis or geospatial analytics, is
concerned with the mining and mapping the locations of social media users,
contents, and data.

Layer Six: Mobile

Mobile analytics is the next frontier in the social business landscape. Mobile
analytics deals with measuring and optimizing user engagement with mobile
applications (or apps for short).

Layer Seven: Hyperlinks

Hyperlink analytics is about extracting, analyzing, and interpreting social media
hyperlinks (e.g., in-links and out-links). Hyperlink analysis can reveal sources of
incoming or outgoing web traffic to and from a web page or website.

Layer Eight: Multimedia

Social media multimedia analytics is the art and science of harnessing business
values from video, images, audio, animations, and interactive content posted over
social media outlets.

5. Types of Social Media Analytics

Social media analytics help achieve business objectives through the describing
of data to analyze trends, predicting future problems and opportunities, and
72 Social Media Analytics in Predicting Consumer Behavior

optimizing business processes to enhance organizational decision making. Like

any kind of business analytics (Delen and Demirkan 2013), social media analytics
can take four forms (Figure 3, Table 3):
1. Descriptive analytics
2. Diagnostic analytics
3. Predictive analytics
4. Prescriptive analytics

Figure 3. Types of social media analytics (Source: Khan 2018)

5.1. Descriptive Analytics (Reactive)

Descriptive analytics is reactive in nature and deals with the questions of “what
happened and/or what is happening?” It is mostly focused on gathering and
describing social media data in the form of reports, visualizations, and clustering
to understand a well-defined business problem or opportunity. Purpose built
or social media embedded platforms (such as Facebook Insights and Twitter
Analytics) dashboards are used to collect and display social media metrics,
such as likes, comments, tweets, and posts. Actions analytics (e.g., no. of likes,
tweets, and views) and text analytics are examples of descriptive analytics. Social
media text (e.g., user comments), for instance, can be used to understand users’
sentiments or identify emerging trends by clustering themes and topics. Currently,
descriptive analytics accounts for the majority of social media analytics.
Introduction to Social Media Analytics 73

Table 3. Business application of social media analytics

Types Example 1 Example 2

Descriptive Twitter Analytics, for example, Facebook Insights allows users
Analytics provides statistics on tweet to track user interaction on
impressions, profile visits, and their Facebook pages. This
audience demographics. It also descriptive analytics tool shows
provides insights on content that statistics, such as, number of
resonates more with your audience, likes, shares, comments, and
measure impressions, engagements, reports on your site and your
and reach. weekly reach.
Diagnostic Can be used to determine the success Diagnostics analytics offers
Analytics of social media marketing campaigns, insights into the behaviour of
for example, to investigate an consumers, such as, a customer
association between a certain searching for specific vehicles
event and reactions to that event through Google, then clicking
on social media to tell you exactly on a landing page of car dealer
where you went wrong. Users are website, but exiting without
then able to utilize this data to finding what they needed.
make improvements to their future
campaigns.
Predictive Social media data can be used to Google used predictive
Analytics predict future events. A study by Asur analytics to collect data
and Huberman (2010) used Tweets regarding the outbreaks of
to forecast box-office revenues for flu. By matching key search
selected movies. By tracking 24 terms associated with people
movies, 2.89 million tweets from 1.2 in different regions of the
million users, they predicted box- world, Google could track flu
office revenues of movies before they outbreaks in near real time.
are released.
Prescriptive Netflix uses large quantities of data Facebook tracks and analyze
Analytics regarding consumer viewing habits users browsing and behavior
to optimize their processes and create data to recommends pages,
recommender systems that provide friends, groups, games that
users with suggestions of content they interest them. Similarly, by
may enjoy based of their previous using review ratings from
viewing habits. like-minded people, it predicts
your likes and interests and
prescribes items it thinks you
will enjoy.

5.2. Diagnostic Analytics (Reactive)

Also reactive in nature, diagnostic analytics deals with the questions of “why
something happened?” Enablers of diagnostics analytics include inferential
statistics, behavioral analytics, correlations, and retrospective analysis, and
outcomes being the cause and effect analysis of business issues. For example, while
74 Social Media Analytics in Predicting Consumer Behavior

descriptive analytics can provide an overview of your social media marketing

campaign’s performances (posts, mentions, followers, fans, page views, reviews,
pins, etc.); diagnostic analytics can distill this data into a single view to see what
worked in your past campaigns and what didn’t.
Social media is often used by consumers to vent frustration over various
products and services. For example, by mining and analyzing social media
posts by the owners of famous automobile brands, such as Hondas, Toyotas, and
Chevrolets, researchers at Virginia Tech Pamplin College of Business were able to
discover vehicle defects across the brands. The data collected could be analyzed
to identify issues relating to the safety and performance of their vehicles and to
better understand the reasons for customer dissatisfaction.

5.3. Predictive Analytics (Proactive)

Predictive analytics involves analyzing large amounts of accumulated social
media data to predict a future event. It is also reactive in nature, and in essence,
deals with the question of “what will happen and/or why will it happen?” Enablers
of predictive analytics include data mining, text mining, Web/media mining, and
statistical time-series forecasting. The primary outcome of predictive modeling
is an accurate projection of the future happenings and the reasoning underlying
such events.
For example, an intention expressed over social media (such as buy, sell,
recommend, quit, desire, or wish) can be mined to predict a future event (such
as a purchase). Alternatively, a business manager can predict sales figures based
on historical visits (or in-links) to a corporate website. The TweepsMap tool, for
example, can help you determine the right time to tweet for maximum alignment
within your audience’s time zones. Alternatively, based on analyzing your social
media users’ languages, it can suggest if it is time to create a new Twitter account
for another language.
A well-known example is the prediction of outbreaks of flu. Google used
predictive analytics to collect data regarding the outbreaks of flu. By matching
key search terms associated with people in different regions of the world, Google
could track flu outbreaks in near real time. This data was compared with traditional
flu surveillance systems and through predictive analytics of the flu season, Google
discovered a correlation with higher search engine traffic for related phrases.
A study conducted by Dublin University in 2011 found that social media data
could be used to predict the outcome of presidential elections. They found that the
number of Tweets associated with election results was the single biggest variable
in predicting the presidential winner. Prior to the 2016 presidential election,
Twitter released an “Election 2016 Candidate Buzz” tracker which did in fact
indicate that President Trump would likely win the election.
Elsewhere, researchers have shown that social media chatter can be used
to forecast box-office revenues for movies, more accurately than those of the
Hollywood Stock Exchange (Asur and Huberman 2010).
Introduction to Social Media Analytics 75

5.4. Prescriptive Analytics (Proactive)

While predictive analytics help to predict the future, prescriptive analytics suggest
the best action to take when handling a scenario (Lustig et al. 2010). For example,
if you have groups of social media users that display certain patterns of buying
behavior, how can you optimize your offering to each group? Like predictive
analytics, prescriptive analytics has not yet found its way into social media data.
The main enablers of prescriptive analytics include optimization and simulation
modeling, multi-criteria decision modeling, expert systems, and group support
systems. Whereas, the main outcome of prescriptive modeling is either the best
course of action when handling multiple scenarios, or expert opinions provided to
a decision maker that could lead to the best possible course of action.
Netflix collects large quantities of data regarding consumer viewing habits
including how long viewers watch, the devices they are using, what time of day
they watch as well as when they paused, rewound or stopped a show. This data
allows Netflix to optimise its processes and create recommender systems that
provide users with suggestions of content they may enjoy based on their previous
viewing habits.
Collaborative filtering is a recommender system technique that assists
consumers in discovering what items are most relevant to them. Within Facebook,
this includes the recommendation of pages, groups, games, and pages. Facebook
collaborative filters bases their recommender system on other people that have
the same interests as you. By using review ratings from like-minded people, it
predicts your likes and interests and prescribes items it thinks you will enjoy.
Finally, predictive/prescriptive analytics models are complex to execute
and have hefty resources. However, in terms of business value it has the largest
potential (Figure 4).

Figure 4. Analytics types and business value (Khan 2018)

76 Social Media Analytics in Predicting Consumer Behavior

6. Social Media Analytics Value Creation Cycle

Social media analytics business value creations are a six step irrelative process
(involving both the science and art) of harnessing the desired business value
from social media data (Figure 5). The value creation journey starts with the
organizational goals and the objectives that we want to achieve with social media
analytics. The business objectives will inform each step of the social media
analytics value creation journey. Business goals are defined at the initial stage,
and the analytics process will continue until the stated business objectives are
fully satisfied. Note that the steps may vary considerably based on the layers of
social media data-mined (and the type of tool employed).

Figure 5. Social media analytics value creation cycle (Khan 2018)

The following are the six general steps, at the highest level of abstraction,
that involves both the science and art of achieving business value from social
media data.
Step 1: Identification
The identification stage is the art part of the social media analytics value creation
process which is concerned with searching for and identifying the right source
of information for analytical purposes. The numbers and types of users and
information (such as text, conversation, and networks) available over social
media are huge, diverse, multilingual, and noisy. Thus, framing the right question
and knowing what data to analyze is extremely crucial in gaining useful business
Introduction to Social Media Analytics 77

insights. The source and type of data to be analyzed should be aligned with
business objectives.
Most of the data for analytics will come from business-owned social media
platforms, such as an official Twitter account, Facebook fan pages, blogs, and
YouTube channels. Some data for analytics, however, will also be harvested from
nonofficial social media platforms, such as Google search engine trends data or
Twitter’s publicly available search stream data. The business objectives that need
to be achieved will play a major role in identifying the sources and type of data
to be mined.
Step 2: Extraction
Once a reliable and mineable source of data is identified, next comes the
extraction of the data. The type (e.g., text, numerical, or network) and size of
data will determine the method and tools suitable for extraction. A small-size of
numerical information, for example, can be extracted manually (e.g., going to
your Facebook fan page and counting likes and copying comments), and a large-
scale automated extraction is done through an API (application programming
interface). Manual data extraction may be practical for small-scale data, but it
is the API-based extraction tools that will help us get the most out of various
social media platforms. Mostly, the social media analytics tools use API-based
data mining.
APIs, in simple words, are sets of routines/protocols that social media service
companies (e.g., Twitter and Facebook) have developed that allow users to access
small portions of data hosted in their databases. The greatest benefit of using
APIs is that it allows other entities (e.g., customers, programmers, and other
organizations) to build apps, widgets, websites, and other tools based on open
social media data. Some data, such as social networks and hyperlink networks,
can only be extracted through specialized tools.
Two important issues to bear in mind here are the privacy and ethical issues
related to mining data from social media platforms. Privacy advocacy groups have
long raised serious concerns regarding the large-scale mining of social media
data and warned against transforming social spaces into behavioral laboratories.
The social media privacy issue first came into the spotlight particularly due to
the large-scale “Facebook Experiment” carried out in 2012. In this experiment,
Facebook manipulated the news feeds feature of thousands of people to see if
emotion contagion occurs without face-to-face interaction (and absence of
nonverbal cues) between people in social networks (Kramer et al. 2014). Though
the experiment was consistent with Facebook’s Data Use Policy (Verma 2014) and
helped to promote our understanding of online social behavior, it does, however,
raise serious concerns regarding obtaining informed consent from participants
and allowing them to opt out.
The bottom line here is that data extraction practices should not violate a
user’s privacy and the data extracted should be handled carefully. The policies
78 Social Media Analytics in Predicting Consumer Behavior

should explicitly detail social media ownership regarding both accounts and
activities such as individual and page profiles, platform content, posting activity,
data handling, and extraction, etc.
Step 3: Cleaning
This step involves removing the unwanted data from the automatically extracted
data. Some data may need cleaning, while other data can go into analysis directly.
In the case of text analytics cleaning, coding, clustering, and filtering may be
needed to get rid of irrelevant textual data using natural language processing
(NPL). Coding and filtering can be performed by machines (i.e., automated) or
can be carried out manually by humans. For example, Discovertext combines
both machine learning and human coding techniques to code, cluster, and classify
social media data (Shulman 2014).
Step 4: Analyzing
At this stage, the clean data is analyzed for business insights. Depending on the
layer of social media analytics under consideration and the tools and algorithm
employed, the steps and approaches to take will greatly vary. For example, nodes
in a social media network can be clustered and visualized in a variety of ways
depending on the algorithm employed. The overall objective at this stage is to
extract meaningful insights without the data losing its integrity.
While most analytics tools lay out a step-by-step procedure to analyze social
data, having background knowledge and an understanding of the tools and their
capabilities are crucial in arriving at the right answers.
Step 5: Visualization
Data or information visualization is the process of converting numerical data into
a graphical (or visual) format to reveal hidden patterns and casual relationships
in data to help facilitate business decision making. In fact, data visualization is
the use of computer-supported, interactive, visual representations of abstract
data to amplify human understanding (Card et al. 1999), thus enabling us to gain
knowledge about the hidden internal structures and causal relationships in data.
The notion of using visuals to understand data and information is not new.
The use of maps and graphs has been around since the 17th century. However, with
the advent of computer programs and affordable tools data visualization is easy
to accomplish with minimum effort and skill. Thanks to power data visualization
tools (such as Tableau and SAS Visual Analytics) anyone can process and visualize
a large amount of data with a click of a button in no time.

7. Visual Analytics
Visual analytics is the science of analytical reasoning facilitated by interactive
visual interfaces (Thomas and Cook 2005). Visual analytics is becoming an
important part of interactive decision making facilitated by solid visualization
Introduction to Social Media Analytics 79

(Wong and Thomas 2004). Effective visualization is particularly helpful with

complex and large data sets because it can reveal hidden patterns, relationships,
and trends. It is the effective visualization of the results that will demonstrate the
value of social media data to top management.
In other words, it is the science and art of employing visualization to analyze
data to facilitate business decision making. According to Keim et al. (2006):
“Visual analytics is more than just visualization and can rather be seen as an
integrated approach combining visualization, human factors and data analysis. …
With respect to the field of visualization, visual analytics integrates methodology
from information analytics, geospatial analytics, and scientific analytics.
Especially human factors (e.g., interaction, cognition, perception, collaboration,
presentation, and dissemination) play a key role in the communication
between human and computer, as well as in the decision making process.”
(Keim et al., 2008).

7.1. Importance of Visualization

The human brain by nature processes visual information easier than numbers and
spreadsheets. Thus, complex ideas and concepts can easily be conveyed using
effective visuals and graphs. Effective visualization plays an important role
in decision making, particularly it is helpful with complex and large data sets
because it can reveal hidden patterns, relationships, and trends. These days visual
communication is a must-have skill for managers (Berinato 2016) because it is
the effective visualization of the results that will demonstrate the value of social
media data to top management.

7.2. Social Media Data Visualization

In addition to numerical results, most of the eight layers of social media analytics
will also result in visual outcomes. Depending on the layer of the analytics, the
analysis part will lead to relevant visualizations for effective communication of
results. Text analytics, for instance, can result in a word co-occurrence cloud;
hyperlink analytics will provide visual hyperlink networks; and location analytics
can produce interactive maps. Depending on the type of data, different types of
visualization are possible, including the following.
Network data (with whom)—Network data visualizations can show who is
connected to whom. For example, a Twitter following-following network chart
can show who is following whom.
Topical data (what)—Topical data visualization is mostly focused on what aspect
of a phenomenon is under investigation. A text cloud generated from social media
comments can show what topics/themes are occurring more frequently in the
discussion.
80 Social Media Analytics in Predicting Consumer Behavior

Temporal data (when)—Temporal data visualization slice and dice data on a time
horizon and can reveal longitudinal trends, patterns, and relationships hidden in
the data. Google Trends data, for example, can visually investigate longitudinal
search engine trends.
Geospatial data (where)—Geospatial data visualization is used to map and locate
data, people, and resources.
One of the interesting data visualization techniques employed is word
cloud. Word clouds are images composed of words, used in a particular text or
subject, where the size of each word indicates its frequency or importance. Many
Text Analytics platforms produce Word Clouds based on an analysis of the text
examined. Wordle is the most well-known word cloud software and is free to use
at Wordle.net. Words are stacked into a box or some other shape with the largest
words being the most prevalent though the information is mostly descriptive.
Other forms of visualizations include trees, hierarchical, multidimensional
(chart, graphs, tag clouds), 3-D (dimension), computer simulation, infographics,
flows, tables, heat maps, plots, etc.
Step 6: Interpretation or Consumption
This step relies on human judgments to interpret valuable knowledge from the
visual data. Meaningful interpretation is of particular importance when we are
dealing with descriptive analytics that leaves room for different interpretations.
While companies are quickly mastering sophisticated analytical methods, skills,
and techniques needed to convert big data into information, there seems to be a
gap between an organization’s capacity to produce analytical results and its ability
to effectively consume it. For example, a study of 2,719 business executives,
managers, and analytics professionals from the world found that the greatest
problem with creating business value from analytics is not data management issues
or complex data modeling skills. But it was translating analytics into business
actions and making business decisions based on the results, not producing the
results per se (Kiron et al. 2014).
The study also reported that there are three levels of analytical maturity in
organizations:
1. Analytically Challenged: These organizations lack sophisticated data
management and analytical skills and generally rely more on management
experience in decision making.
2. Analytical Practitioners: Such organizations tend to use analytics for
operational purposes, have “just good enough data” and are working to
become more data driven.
3. Analytical Innovators: Analytical innovators organizations are more strategic
in their use and application of analytics, place greater value on data, and
Introduction to Social Media Analytics 81

have higher levels of data management and analytical skills. These are the
organizations that are the most successful in translating analytics results into
business actions and decisions making.

Improving Analytics Consumption Abilities

Having domain knowledge and expertise is crucial in consuming the obtained
results correctly. To improve such abilities, the following strategies or approaches
can be employed.
1. Building an analytics vocabulary—While for most managers it is not
necessary to understand advanced analytics topics and techniques, it is crucial
to build an analytics vocabulary and get familiar with the basic concepts such
as statistics, machine learning, data management and big data, descriptive,
predictive, and prescriptive analytics.
2. Producing easily consumable results—This approach requires training data
scientists and analysts to produce interactive and easy-to-use visual results.
Managers need to talk to their data scientists to create results that they are
comfortable consuming.
3. Improving consumption capabilities—This strategy focuses on improving
management analytics consumption capabilities.
4. Recognizing the limitations of the analytical model—Analytical models
and algorithms are not perfect are very sensitive to variation in data inputs.
For effective decision making, analytical models should be complemented
with management’s knowledge of the changing business context. This will
help identify the limitation of analytical models and commission additional
analysis to understand the potential effects of variables not covered by
analytical models.

8. Current VS., Potential Use of Social Media

Analytics
As mentioned earlier, at present, the majority of the analytics industry and practice
revolves around descriptive analytics. And if there is any use of predictive and
prescriptive analytics, it is limited to structured data only. According to Garner
(2013), only 3 percent of companies used prescriptive analytics, but only with
structured data. However, the use of social media data for descriptive analytics
is just the tip of the iceberg. Its true potential is in predictive and prescriptive
analytics (see Figure 6). The future of the analytics industry is in the use of
predictive/prescriptive analytics thus unleashing the true potential of social
media analytics.
82 Social Media Analytics in Predicting Consumer Behavior

Figure 6. Current VS., potential use of social media analytics (Khan 2018)

9. Challenges to Social Media Analytics

Social media data is high volume, high velocity, and highly diverse, which, in
a sense, is a blessing regarding the insights it carries; however, analyzing and
interpreting it presents several challenges. Analyzing unstructured data requires
new metrics, tools, and capabilities, particularly for the real-time analytics that
most businesses do not possess. Some social media analytics tools are listed in a
later section.

9.1. Volume and Velocity as a Challenge

Social media data is large and generated swiftly. Capturing and analyzing millions
of records that appear every second is a real challenge. For example, on Twitter,
three hundred forty-two thousand tweets appear every minute, and on Facebook,
one million likes are shared every twenty minutes. Capturing all this information
may not be feasible. Knowing what to focus on is crucial for narrowing down the
scope and size of the data. Luckily, sophisticated tools are being developed to
handle high-volume and high-velocity data.

9.2. Diversity as Challenge

Social media users and the content they generate are extremely diverse,
multilingual, and vary across time and space. Not every tweet, like, or user is
worth looking at. A tweet or mention coming from an influential social media
user is more valuable than a tweet from a non-influential user. Due to the noisy
and diverse nature of social media data, separating relevant content from noise is
challenging and time-consuming.
Introduction to Social Media Analytics 83

9.3. Unstructured Data as a Challenge

Unlike the data stored in corporate databases, which are mostly numbers, social
media data is highly unstructured and consists of text, graphics, actions, and
relations. Short social media text, such as tweets and comments, has dubious
grammatical structure, and is laden with abbreviations, acronyms, and emoticons
(a symbol or combination of symbols used to convey emotional expressions in
text messages), thus representing a significant challenge for extracting business
intelligence.

10. Social Media Analytics Accuracy

Owing to the challenges of volume, velocity, and diversity, the accuracy of Social
Media Analytics is questionable. As the huge unstructured data (also known as
‘dirty’ data) is generated over social media, the accuracy of social listening is
decreasing. For example, text analytics (one of the main components of social
media analytics) cannot capture many of the ways people use language. Most of
the tools have been developed by Western and English-speaking countries. The
tools often translate the text into English, apply sentiment analysis, then assign
it to the original post in its native language. This approach is problematic; if
sentiment analysis is not available in the native language, it should not be offered
in the first place. One has but to look at translations done by automated tools
such as Babelfish to see just how mangled language translations become when
processed in this manner. Similarly, many users are posting images and videos
instead of text; most of this will be invisible to text based social media analytics
platforms.
In addition, as of 2016, social media analytics platforms have not yet tapped
into most of the new AI capabilities. There is much room for growth in this area.
Hybrid Systems (usually boutique) have been developed for specialized use cases
where encoded visual, auditory, and hepatic (i.e., touch sensor) data is used to
provide context to Social Media, such that information is usually not present
in verbatim.

11. Social Media Analytics Industry

The social media analytics industry is huge and is growing rapidly. According to
some estimates, the global social media analytics market has grown at a Compound
Annual Growth Rate of 27.6% (MarketsandMarkets 2016). It is predicted that the
total value of the social media analytics market will be USD 41.38 billion by
2029 as compared to USD 9.26 billion in 2022. The primary contributing growth
factors of social media analytics market are the growing number of social media
users, greater spending on analytics, and more focus on market and competitive
intelligence.
84 Social Media Analytics in Predicting Consumer Behavior

The social media ecosystem is complex and comprises several players

and industries. Figure 7 provides an overview of the social media ecosystem,
containing a diversity of players including:
• Analytics vendors
• Analytics service providers
• System integrators
• Mobile application providers
• Consulting service providers
• IT service providers
• Resellers
• Telecom operators
• Enterprise users, and
• Technology providers.

Figure 7. Social media ecosystem

12. Social Media Analytics Tools

Social media analytical tools are also coming to market at a high pace to keep
up with the growing need for analyzing the vast amount of data. Social media
analytics tools come in a variety of forms and functionalities. Mainly it can be
divided into: (1) analytical applications that do not require programming skills,
and (2) tools/scripts/modules that require programming skills.
Introduction to Social Media Analytics 85

Table 4 lists some example tools on each layer of social media analytics. These
tools can be used to measure different layers of social media data, especially when
aligned with an organization’s business strategy. The Digital Methods Initiative
(DMI) Internet Studies research group provides a compresence list of methods
and tools for social media data (the list can be accessed using this shortening
URL: goo.gl/EiTWi

Table 4. Examples of social media analytics tools with respect to its layers

Data Tools
Text Discovertext
Lexalytics
Tweet Archivist
Twitonomy
Netlytic
LIWC
Voyant
Actions Lithium
Twitonomy
Google Analytics
SocialMediaMineR
Network NodeXL
UCINET
Pajek
Netminer
Flocker
Netlytic
Reach
Mentionmapp
Mobile Countly
Mixpanel
Google Mobile Analytics
Location Google Fusion Table
TweepsMap
Trendsmap
Followerwonk
Esri Maps
Agos
(Contd.)
86 Social Media Analytics in Predicting Consumer Behavior

Table 4. (Contd.)

Data Tools
Hyperlinks Webometrics Analyst
VOSON
Research Engines Google Trends
Google Correlate
Multimedia Crimsonhexagon Image Analytics
YouTube Analytics
SAS Visual Analytics
Google Cloud Vision API
Simply Measured

References
Asur, S. and B.A. Huberman. 2010. Predicting the future with social media. 2010 IEEE/
WIC/ACM International Conference on Web Intelligence and Intelligent Agent
Technology, pp. 492–499: 10.1109/WI-IAT.2010.63
Bekmamedova, N. and G. Shanks. 2014. Social media analytics and business value: A
theoretical framework and case study. 2014 47th Hawaii International Conference on
System Sciences, pp. 3728–3737: 10.1109/HICSS.2014.464
Berinato, S. 2016. Visualizations that really work. Harvard Business Review (June 2016).
Card, S., J. Mackinlay and B. Shneiderman. 1999. Readings in Information Visualization:
Using Vision to Think. Morgan Kaufmann Publishers.
Chen, H., R.H.L. Chiang and V.C. Storey. 2012. Business intelligence and analytics: From
big data to big impact. MIS Q., 36(4), 1165–1188.
Delen, D. and H. Demirkan. 2013. Data, information and analytics as services. Decision
Support Systems, 55(1), 359–363. http://dx.doi.org/10.1016/j.dss.2012.05.044
Keim, D., G. Andrienko, J-D. Fekete, C. Görg, J. Kohlhammer and G. Melancon. 2008.
Visual analytics: Definition, process, and challenges. pp. 154–175. In: J.T.S. Andreas
Kerren, Jean-Daniel Fekete and Chris North [Eds.]. Information Visualization. Vol.
4950, Berlin, Heidelberg Springer-Verlag.
Khan, G.F. 2013. Social media-based systems: An emerging area of information systems
research and practice. Scientometrics, 95(1), 159–180. 10.1007/s11192-012-0831-5
Khan, G.F. 2018. Creating Value with Social Media Analytics: Managing, Aligning, and
Mining Social Media Text, Networks, Actions, Location, Apps, Hyperlinks, Multimedia,
Search Engines Data. CreateSpace Independent Publishing Platform, April 23, 2018.
Kiron, D., P.K. Prentice and R.B. Ferguson. 2014. Raising the bar with analytics. MIT
Sloan Management Review, https://sloanreview.mit.edu/article/raising-the-bar-with-
analytics/
Kramer, A.D.I., J.E. Guillory and J.T. Hancock. 2014. Experimental evidence of massive-
scale emotional contagion through social networks. Proceedings of the National
Academy of Sciences, 111(24), 8788–8790. 10.1073/pnas.1320040111
Introduction to Social Media Analytics 87

Lustig, I., B. Dietrich, C. Johnson and C. Dziekan. 2010. The analytics journey: An IBM
view of the structured data analysis landscape: Descriptive, predictive and prescriptive
analytics. Analytics-Magazine, available at: http://www.analytics-magazine.org/
november-december-2010/54-the-analytics-journey.
MarketsandMarkets. 2016. Social Media Analytics Market worth 9.54 Billion USD by
2022. https://www.marketsandmarkets.com/PressReleases/social-media-analytics.asp:
Petersen, R. 2012. 166 Cases Studies Prove Social Media Marketing ROI. BarnRaisers.
Syncapse. (2013). THE VALUE OF A FACEBOOK FAN 2013: Revisiting Consumer Brand
Currency in Social Media. New York, NY.
Thomas, J.J. and K.A. Cook. 2005. Illuminating the Path: The Research and Development
Agenda for Visual Analytics. IEEE Press.
Verma, I.M. 2014. Editorial Expression of Concern: Experimental evidence of massive
scale emotional contagion through social networks. Proceedings of the National
Academy of Sciences, 111(29), 10779. 10.1073/pnas.1412469111
Wong, P.C. and J. Thomas. 2004. Visual Analytics. IEEE Computer Graphics &
Applications, 24, 20–21.