Fabricated information on social media: A Case Study of Twitter

Abstract

By increasing the accessibility to social media, such as Twitter, Facebook and Google+, from various devices it is a challenge to recognize fabricated or false information. This misinformation might be created by the wrong interpretation or translation of a subject, which might affect its core significance. Recent research shows that most people are not aware of the distribution of fabricated and false information on social media, a problem that is also evident in respect of information pertaining to Islam. Consequently, there are various people who are sharing false interpretations and wrong translations of Islamic hadiths on social media. In this work, we shall be developing a system that will be able to extract recent tweets about Islamic hadiths from Twitter and investigate their authentication by comparing them with a reliable database. Therefore, we will develop a database system that will be able to store all the authenticated Islamic hadiths with their Malay language translation. This system will also be able to extract recent tweets about Islamic hadiths from Twitter via the Twitter API functions. Additionally, from a comparison of the extracted tweets and al-hadith stored on our database, it will be possible to analyze the percentage of fabricated tweets about al-hadiths in the region of Malaysia. The development of this system will be completed when all the data are stored appropriately. The extraction of relevant tweets from Twitter will be achieved, and, simultaneously, the authenticity of tweets will be investigated.

Full text

Fabricated information on social media:
A Case Study of Twitter
Ali Ebrahimi1, Maizatul Akmar Ismail1, Salinah Jaafar2, Suad Awab3, Rohana Mahmud1, Abdullah
Ghani1, Ishak Suliaman4, Amirrudin Kamsin1, Norisma Idris1
1Faculty of Computer Science and Information Technology;
2Academy Of Malay studies
3Faculty of Language and Linguistics
4Academy of Islamic Studies
University of Malaya,
50603 Kuala Lumpur, Malaysia.
maizatul@um.edu.my
Abstract: By increasing the accessibility to social media, such as
Twitter, Facebook and Google+, from various devices it is a
challenge to recognize fabricated or false information. This
misinformation might be created by the wrong interpretation or
translation of a subject, which might affect its core significance.
Recent research shows that most people are not aware of the
distribution of fabricated and false information on social media, a
problem that is also evident in respect of information pertaining to
Islam. Consequently, there are various people who are sharing false
interpretations and wrong translations of Islamic hadiths on social
media. In this work, we shall be developing a system that will be
able to extract recent tweets about Islamic hadiths from Twitter and
investigate their authentication by comparing them with a reliable
database. Therefore, we will develop a database system that will be
able to store all the authenticated Islamic hadiths with their Malay
language translation. This system will also be able to extract recent
tweets about Islamic hadiths from Twitter via the Twitter API
functions. Additionally, from a comparison of the extracted tweets
and al-hadith stored on our database, it will be possible to analyze
the percentage of fabricated tweets about al-hadiths in the region of
Malaysia. The development of this system will be completed when
all the data are stored appropriately. The extraction of relevant
tweets from Twitter will be achieved, and, simultaneously, the
authenticity of tweets will be investigated.
Keywords: Data Extraction, Data Analysis, Online Social Media,
Twitter, Fabricated Information
1. Introduction
In the past few years, online social media (OSM), which
encourages people to generate a particular network to share
information, opinions, and thoughts in the form of text,
photos and videos, has emerged as one the most popular
phenomenon on the Web. This scenario is an advantage of
Web 2.0, which shifted users from being passive consumers
of information to data producers. Thus, OSM platforms
provide novel and unique research opportunities; for
instance, the analysis of a large scale of user interaction
provides a chance to answer questions concerning how
humans indulge in friendships and how these friendships
grow over time [1]. In addition, human activity and shared
opinion in OSM may also benefit scientists and enable them
to understand human behavior and identify groups of users
that share information on OSM without necessary reviewing
a certain topic. Therefore, in this study, we are planning to
use social media to investigate an OSM in order to determine
the number of users who share fabricated information.
Web data extraction systems have become powerful
technologies that are used extensively for applications in the
analysis process of web documents in full text format for a
wide range of usage in business intelligence [2-5]. Such uses
include medical [6, 7] and company management, to
understand human behavior [8-10], bio-informatics [11],
tourism and travel [12], affectivity of media on politics [13]
and crawling of social web platforms [14, 15]. However,
there is insufficient research about fabricated data and the
information that is shared on OSM platforms.
Several OSMs empower their users to communicate with
their network via a variety of functions. A brief study on
popular social media – Facebook[16], Twitter[17] and
LinkedIn [18] – shows that the most common function
among them is the Share function which is called “Tweet”
on Twitter, “Post” on Facebook, and “Share” on LinkedIn.
This function allows users to share their opinion or
information via different formats like texts, images, videos
and sounds with additional abilities like tagging a contact or
attaching location and emoticons to their posts. This function
brings a huge amount of public data that has the potential to
deliver valuable information to the scientific community,
marketing agencies, NGOs and other organizations interested
in people’s behavior and their points of view [19].
As the number of registered users on OSMs is rapidly
growing, and, at the same time, the size of content shared is
rising, there is a great opportunity for scientists to extract
valuable information about human behavior and social
interactions. Twitter is an OSM that can be utilized as a tool
for research. Currently it has over 284 million monthly
active users that generate 500 million tweets per day [20].
The advantage of Twitter in comparison to other social
media is that Twitter is an open platform, which allows
researchers to access tweets by using the Twitter application
programming interface (API) [21]. This presents three APIs
including SEARCH API, STREAM API and REST API.
These functions offer an opportunity for researchers to
access a large quantity of data [22].
In this project, our main objective is to develop a system that
will be able to extract recent tweets from the Twitter website
regarding Islamic hadiths. This information will then be
compared to a database that stores all Islamic hadiths in the
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)
204

Malaysian language [23]. Moreover, these compared tweets
will be analyzed to study the percentage of fabricated
information that people are tweeting on the Twitter website.
The Islamic world is growing at a faster rate than ever
before, which has led to various Islamic sects and
perspectives. Simultaneously, the rapid rise of using social
media for sharing information on the World Wide Web has
caused an increase in the amount of fabricated data
concerning the translation and interpretation of Islamic
hadiths. As discussed earlier, some OSMs allow developers
to gather public information. For instance, Twitter presents
API functions that enable developers to extract recent tweets
for different aspects.
Nowadays, identifying the accuracy of Islamic hadiths and
the interpretation and authenticity thereof is a problem that
Muslims face, and Muslim scholars are consistently working
hard to impede the spread of false, inappropriate
transliterations and interpretations. However, it is difficult
for them to prevent the spread of these problems in
cyberspace, especially on social media. Therefore,
developing an application that might be able to gather
information from social media can provide a solution for the
above-mentioned problems. Such an application will create a
base on the Twitter API enabling us to extract recent tweets
from the Twitter website.
The crux of this project is to develop a system to extract the
recent tweets (specified tweets) from Twitter and store them
in the form of structured data on a database to prepare them
in order to make a comparison between the tweet and an
Islamic hadith database that was previously developed by
Bimba et al. [23].
The core objectives of this project are to review the available
research, methods and architectures of data extraction from
social media that have been carried out by other scholars and
researchers in this field. In addition, this study intends to
design and develop a system architecture that will be able to
extract recent tweets about Islamic hadiths in the region of
Malaysia from the Twitter website and store them on an
appropriate database. Moreover, extracted tweets will be
analyzed to determine the amount of fabricated information
about Islamic hadiths that is shared on Twitter.
This paper is organized as follows: Section 2 presents a brief
review of the use of Twitter in other areas of research;
Section 3 presents the tools and toolsets that will be used in
this project; in Section 4, the challenges of OSM data
extraction and the advantages of OSM data extraction are
discussed.
2. Related work
The theme of Twitter’s Streaming API has been referenced
by a number of reviews. Some researchers used Twitter’s
Streaming API to generate systems that enable scientists to
analyze the behavior of the users of Twitter while they are
using social media and online networks. In addition, [19]
developed an architecture for a Twitter data collection. There
are several studies that show that Twitter has been used to
collect data for content modeling [24, 25] and statistical
analysis of content [26].
Nevertheless, because of the popular usage of Twitter’s
Streaming API in several scientific fields, it is necessary that
we understand how having a collection of information or
data affects our final results. For instance, many researchers
have used Twitter to detect real time social and physical
events. [27] investigated the usage of Twitter as a device to
detect frequent and diverse social and physical events in real-
time; they examined the 2010-2011 US National Football
League (NFL) games. [28,29] also studied a real time sport
event. [30, 31] examined the detection of earthquakes, and
[32] studied events pertaining to natural hazards (grassfires
and floods). [33] examined recent news topics and bunched
the corresponding tweets into breaking news. In terms of
political viewpoints, researchers used Twitter to determine
people’s reactions to the media [13].
One of the advantages of real time detection is that the
collected data enables scientists to analyze the behavior of
people at the time that they tweet their opinion on social
media. As mentioned previously, the aim of this research is
to identify fabricated information on Twitter, which requires
the extraction of data in real time. Thus, our work mimics
cases that detect real time events. The other aspect that we
have to consider is the accuracy of the collected data the
same as other researchers with similar objectives.
Figure 1. Architecture of Twitter extraction
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)
205

Figure 2. Evaluation of extracted tweets
3. System development
Twitter is a microblogging system that enables users to post
and receive short messages of up to 140 characters, which
are called "tweets". The actual number of the characters is
160 based on the size defined for the Short Message Service.
However, this limit is split into 140 characters for the post-
box and another 20 characters for the username. On the other
hand, this limitation motivates users to produce creative
posts. Twitter allows users to share URLs that may contain
Videos, articles and any other information that the tweeter
would like to share with his followers by Twitter’s short
messaging service, which is called twl.tl URL. This poses an
additional difficulty for scientists if the URL is a part of the
analysis system. Twitter offers considerable functionality to
users for posting a simple message including the ability to
send a message or reply to a friend. The message can be
addressed to a specific user either on their follower list or by
starting the message with “@Username”. The other
functionality of Twitter is the ‘hashtag’, which introduced
users to tagging their post with a word, number or phrase in
order to make it easier for people to find and follow
discussions about a specific subject.
Twitter offers three application programming interfaces
(APIs): Rest API, Search API and Streaming API. The Rest
API allows developers to read and write tweets by
programmatic access. This means that it allows developers to
access Twitter data and store it in a database; the Search API
returns recent or popular tweets that match with a specific
query, and the Streaming API provides real-time access to
Twitter data. Therefore, these APIs allow developers to
achieve their target when real time data analysis is used for
scientific reasons.
Fig. 1 shows a simplified architecture proposed for tweet
collection. The process starts by choosing the right API and
applying a code to the system and then storing tweets on a
database as unstructured data. The initial step for gathering
tweets will be done with the help of the PHP programing
language, which will run as a continuous background
process. When a user sends a tweet in real time, the Twitter
streaming API will employ the PHP query to insert it into the
MySQL table. Afterwards the unstructured data may be
normalized and stored in a database in order to be used for
future reference.
After collecting data from Twitter and storing it as structured
data, it will be evaluated. In this process, the database of
collected tweets will be compared with al-hadith data by
applying a text similarity algorithm that is responsible for
determining the similarity between texts. This means that the
algorithm will be applied to identify the most comparable al-
hadith in the database and compare it to the extracted tweets.
This will display the percentage of similarity of the tweets
based on the al-hadith database. Figure 2 illustrates the
overall evaluation process. Raw unevaluated tweets will be
sieved by the extracted tweets database and al-hadith
database. The text similarity algorithm will be applied to
both these databases and will eventually provide true
percentages of tweets when compared to the authenticated
al-hadith database.
In this project, to actualize the above-mentioned tweet
collection architecture, Twitter APIs are going to be
developed in conjunction with the PHP programming
language and MySQL. More specifically, the first step of
data collection will be achieved using the PHP programming
language, which is a continuously running program that
works to extract new related tweets about the assigned
subjects. The extracted tweets will then be inserted into the
MySQL table. After collecting data, a text similar algorithm
will be applied in order to do the evaluation process.
4. Discussion
4.1. Challenges of Social Media Extraction:
Although information extraction from OSM and data
analysis has been widely used by many organizations that
analyze data for different fields, research on web data
extraction shows there are still many critical challenges that
may affect the quality of web data extraction. Typically, the
problems in extracting data from the Web are challenging
because they require several tools and techniques that are not
easily accessible. In this section, the system development,
velocity of extraction, availability of data and data
management system are elaborated upon.
 System development:
To extract raw data from an OSM like Twitter, developers
often create various system that are able to extract the
requested data. This means that the best way to create a
system for extraction is to divide the data in the extraction
program into many modules, and, subsequently, command
them to extract a part of the data. This data is then combined
together as a dataset. Such a process requires an expert
programmer who is able to face this challenge.
 Velocity of extraction:
One of the main targets of data extraction from OSMs is to
prepare a set of information either for analysis in exclusive
areas, such as businesses, or create a collection of vital
information for organizations like the government. For
instance, in the field of business intelligence, analyzing data
quickly and efficiently can provide a competitive advantage.
It is a strength for a company to have the first set of data in
order to plan their strategy at the appropriate time. Therefore,
the tools and techniques that are used for data extraction
must be able to mine a large amount of data in the shortest
possible time.
Status updates on social media also raise many practical
performance challenges in that data extraction needs to be
performed in the run time. A huge amount of information
must be extracted within an appropriate time, which
exacerbates the difficulty of the technical aspects of data
extraction.
 Availability of data:
Information on the Web is heterogeneous. Many users of
OSM may provide the same or similar information using
different words and formats. This problem raises the
challenge of choosing the right source for the data extraction
process, which requires the extraction source that presents
the most accurate information.
 Data management system:
To be more efficient and valuable, the data extracted from
OSM should be stored on a database. The rapid growth of
OSM data may cause an increase in the volume of the
extracted data, which may be called Big Data. Big Data
refers to “massive data sets” with a large and complex
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)
206

structure that requires a high capacity storage database and
an architecture to manage it. For these kinds of system,
applications must have a strong management system in order
to categorize the extracted data and create integration
between other applications for future actions.
4.2. Advantages of Twitter for Data Extraction
For OSM data extraction, Twitter is a great tool. The key to
its popularity and broad appeal to people lies in its innate
openness for community consumption. It also provides high
performance tooling for developers and researches because it
offers thorough and well-organized data and information.
Twitter data is especially remarkable because it is available
to its users as soon as it is uploaded. Hence, this multifaceted
application represents the true cross-section of the
community it serves. As mentioned previously, Twitter
“following” interactions can range from small dialogues to
interest groups that cater to the information for which people
have an affinity. Another important advantage of Twitter
data is that each tweet has its own author. To perform
sentiment analysis, we need to label data before applying
suitable learning algorithms. For example, [34, 35] found
that for sentiment analysis, the training data needs to be
stored and Twitter has the advantage that it possesses
tweeter-provided sentiment indicators.
5. Conclusion
In conclusion, this case study will demonstrate that
Twitter is a viable tool for data extraction and analysis
for mass consumption and research in Malaysia. This
study will attempt to pool data from Twitter and
evaluate its authenticity since people can freely state
what they had in mind on OSMs without reviewing
their comments. It is imperative to a have type of
'information police' so that only true and authentic
information is allowed to appear on OSM. Further
studies could incorporate Twitter as a device for data
collection, extraction and analysis from OSM because
of Twitter’s broad appeal, user compatibility and API
tools for developers.
References
[1] J. Kleinberg, The convergence of social and
technological networks. Communications of the ACM,
2008. 51(11): p. 66.
[2] A. Mikroyannidis, B. Theodoulidis, and A. Persidis,
PARMENIDES: towards business intelligence discovery
from web data. in Web Intelligence, 2006. WI 2006.
IEEE/WIC/ACM International Conference on. 2006.
IEEE.
[3] R. Baumgartner, et al. Web data extraction for business
intelligence: the lixto approach. 2005: na.
[4] W. F. Cody, et al. The integration of business intelligence
and knowledge management. IBM systems journal, 2002.
41(4): p. 697-713.
[5] J. Srivastava, R. Cooley, Web business intelligence:
Mining the web for actionable knowledge. INFORMS
Journal on Computing, 2003. 15(2): p. 191-207.
[6] K. Denecke, W. Nejdl, How valuable is medical social
media data? Content analysis of the medical web.
Information Sciences, 2009. 179(12): p. 1870-1880.
[7] G. Eysenbach, et al. Empirical studies assessing the
quality of health information for consumers on the world
wide web: a systematic review. Jama, 2002. 287(20): p.
2691-2700.
[8] J. Kleinberg, The small-world phenomenon: An
algorithmic perspective. in Proceedings of the thirty-
second annual ACM symposium on Theory of computing.
2000. ACM.
[9] L. Backstrom, et al. Four degrees of separation. in
Proceedings of the 4th Annual ACM Web Science
Conference. 2012. ACM.
[10] M. E. Newman, The structure and function of complex
networks. SIAM review, 2003. 45(2): p. 167-256.
[11] C. Plake, et al. AliBaba: PubMed as a graph.
Bioinformatics, 2006. 22(19): p. 2444-5.
[12] Z. Xiang, U. Gretzel, Role of social media in online
travel information search. Tourism Management, 2010.
31(2): p. 179-188.
[13] D. A. Shamma, L. Kennedy, and E. F. Churchill. Tweet
the debates: understanding community annotation of
uncollected sources. in Proceedings of the first SIGMM
workshop on Social media. 2009. ACM.
[14] M. Gjoka, et al. Walking in Facebook: A case study of
unbiased sampling of OSNs. in INFOCOM, 2010
Proceedings IEEE. 2010. IEEE.
[15] S. A. Catanese, et al. Crawling facebook for social
network analysis purposes. in Proceedings of the
international conference on web intelligence, mining
and semantics. 2011. ACM.
[16] Facebook. Available from: http://www.facebook.com.
[17] Twitter. Available from: http://www.twitter.com/.
[18] Linkedin. Available from: https://www.linkedin.com/.
[19] M. Oussalah, et al. A software architecture for Twitter
collection, search and geolocation services.
Knowledge-Based Systems, 2013. 37: p. 105-120.
[20] Online source. Available from:
http://www.about.twitter.com/company.
[21] Online documentation. Available from:
https://dev.twitter.com/overview/documentation.
[22] B. K. Chae, Insights from Hashtag# SupplyChain and
Twitter Analytics: Considering Twitter and Twitter
Data for Supply Chain Practice and Research.
International Journal of Production Economics, 2015.
[23] A. Bimba, et al. Towards Enhancing the Compilation of
Al-Hadith Text in Malay. in International Conference
on World Islamic Studies 2015. Seoul, South Korea.
[24] L. Hong , B. D. Davison, Empirical study of topic
modeling in twitter. in Proceedings of the First
Workshop on Social Media Analytics. 2010. ACM.
[25] A. Pozdnoukhov, C. Kaiser, Space-time dynamics of
topics in streaming text. in Proceedings of the 3rd ACM
SIGSPATIAL international workshop on location-based
social networks. 2011. ACM.
[26] M. Mathioudakis, N. Koudas. Twittermonitor: trend
detection over the twitter stream. in Proceedings of the
2010 ACM SIGMOD International Conference on
Management of data. 2010. ACM.
[27] S. Zhao, et al. Human as real-time sensors of social
and physical events: A case study of twitter and sports
games. arXiv preprint arXiv:1106.4300, 2011.
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)
207

[28] D. Chakrabarti, K. Punera, Event Summarization Using
Tweets. ICWSM, 2011. 11: p. 66-73.
[29] J. Hannon, et al. Personalized and automatic social
summarization of events in video. 2011: p. 335.
[30] T. Sakaki, M. Okazaki, and Y. Matsuo, Earthquake
shakes Twitter users: real-time event detection by
social sensors, in Proceedings of the 19th international
conference on World wide web. 2010, ACM: Raleigh,
North Carolina, USA. p. 851-860.
[31] Y. Qu, et al. Microblogging after a major disaster in
China: a case study of the 2010 Yushu earthquake. in
Proceedings of the ACM 2011 conference on Computer
supported cooperative work. 2011. ACM.
[32] S. Vieweg, et al. Microblogging during two natural
hazards events: what twitter may contribute to
situational awareness. in Proceedings of the SIGCHI
conference on human factors in computing systems.
2010. ACM.
[33] J. Sankaranarayanan, et al. Twitterstand: news in
tweets. in Proceedings of the 17th acm sigspatial
international conference on advances in geographic
information systems. 2009. ACM.
[34] A. Bifet, E. Frank. Sentiment knowledge discovery in
twitter streaming data. in Discovery Science. 2010.
Springer.
[35] M. A. Russell, Mining the Social Web: Data Mining
Facebook, Twitter, LinkedIn, Google+, GitHub, and
More. 2013: " O'Reilly Media, Inc.".
The 4th International Conference on Computer Science and Computational Mathematics (ICCSCM 2015)
208