IEEE Technology and Society Magazine - June 2018 - 47

identified topics. Topics related to hijab and corresponding business goals, were infrequent, and rarely
retweeted. We also calculated the impression to identify the most influential users, such as religious leaders, news agencies, and journalists.

Utility of Twitter for Investigating
Social Phenomena
Online social media, such as the micro-blogging site
Twitter, have become a rich source of real-time data
representative of online human behaviors [1]. Unless
the Twitter users mark tweets as private, tweets are
public, providing information from a broad range of
people on a variety of topics [2]. For example, Twitter is
used for sharing information about healthcare, food
consumption, social events, crisis, political views, sports,
and culture.
Larsson and Moe [3] utilized online tools and presented a rationale for data collection and analysis of
Twitter users, during the 2010 Swedish national election. Veenstra et al. [4] investigated Twitter use in citizen
journalism in the 2011 Wisconsin labor protests. In this
study, 775 030 tweets revealed differences in mobile
and non-mobile use. Mobile users, who may have been
present at the protests, posted fewer URLs overall; however, when they did post URLs, they were more likely
to link to traditional news sources and to provide additional hashtags for context. Over time, all link posting
declined, as users became better able to convey firsthand information.
In yet another study, collection of tweets posted from
all local health departments (LHDs) having a Twitter
account identified tweets related to the subject of diabetes in 2012 [5]. Using content analysis, the researchers
grouped the diabetes-related tweets into categories and
monitored LHDs' programs and trends. Ceron [6] used
Twitter to predict Italian 2011 elections, French national
2012 elections, and the subsequent French legislative
election. While Internet users are not necessarily representative of the whole population of a country's citizens,
the Ceron analysis shows a remarkable ability for social
media to forecast electoral results. In 2014 specific
samples from the Twitter followers of the Canadian
Football League led to insights about what motivates
and satisfies Twitter followers of particular professional
sport teams [7]. Selim et al., [8] coded and analyzed
about 5000 tweets of users from Saudi Arabia and the
United Kingdom, to explore identity motives on Twitter.
Their findings suggest that Saudi users appeared to
seek the socio-psychological value of "distinctiveness,"
whereas British users appeared to seek out "belonging."
In research conducted by Abbar et al. [9], the investigators examined the potential of Twitter to provide insight
into U.S. dietary choices, by linking the tweeted dining
JUNE 2018

∕

experiences of 210 000 users to their interests, demographics, and social networks. Relating the caloric values of the foods mentioned in the tweets to state-wide
obesity rates resulted in a correlation across the (fifty)
50 continental United States including the District of
Columbia. The authors built a model to predict countywide obesity and diabetes statistics, based on a combination of demographic variables and food names
mentioned on Twitter.

Subject of Current Research
A topic of controversy in predominately Islamic countries surrounds women and the wearing of the hijab. A
hijab is a veil that covers the head and chest that is
worn by certain Muslim women beyond the age of
puberty, in the presence of adult males outside of their
immediate family, as a form of modest attire. The hijab
can further denote any head, face, or body covering
worn by Muslim women that similarly conforms to a certain standard of modesty [10].
The purpose of this research is to investigate the
manner in which Twitter is employed in the context of
issues regarding the use or wearing of hijab. In particular we endeavored to determine:
1) which segments of the population most tweet
about matters concerning the wearing of the hijab; 2)
categories identifying the main topics of veil-related
tweets; 3) the degree to which Twitter users react to
hijab commercial advertisement; 4) which groups of
Twitter users influence the release of veil information.

Methods
The study represented was cross-sectional and descriptive. We sampled messages using the network analysis
tool NodeXL [11] to collect tweets related to the keyword
," which has
"hijab," symbolized by the character "
identical meaning in three languages: Arabic, Persian,
and Urdu. We used percent encoding, according to
RFC3986 [12] to supply our search term in NodeXL, as
it does not support unicode characters as input. This
resulted in 10 592 tweets, from which irrelevant messages were removed for a population of 6046 tweets for one
week from August 28 to September 3, 2015. Using Language Detection API [13], the language of all tweets was
detected. Figure 1 illustrates different language engagements compared with populations of people speaking in
those languages in the world. Populations were extracted
from Wikipedia [14], [15] and presented in Table 1.
Excluding Urdu tweets, Arabic and Persian tweets were
coded into 11 different groups. Each group implies a
unique topic. These topics are displayed in Table 2. Figure 2 illustrates frequencies of tweets in each topic.
We identified the most influential users by calculating impressions [16], meaning the number of Twitter

IEEE TEchnology and SocIETy MagazInE

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - June 2018