Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
IDENTIFICATION OF PUBLIC SENTIMENT OVER
COMMENTS THROUGH TWEETS BY DIGITAL INDIA
Dr. Amrita Chakraborty
Assistant Professor, Department of Languages, Literature and Aesthetics (LLA), School
of Liberal Studies (SLS), Pandit Deendayal Petroleum University (PDPU), Gandhinagar
Official Address: Dr. Amrita Chakraborty, Assistant Professor, Department of
Languages, Literature and Aesthetics (LLA), School of Liberal Studies (SLS), Pandit
Deendayal Petroleum University (PDPU), Knowledge Corridor, Raisan Village,
Gandhinagar – 382007, Gujarat, INDIA
Author’s Profile: Dr. Amrita Chakraborty is currently working as an Assistant
Professor with the School of Liberal Studies (SLS) under Pandit Deendayal Petroleum
University (PDPU) in Gandhinagar. A UGC-NET qualified scholar, she has a total
experience of more than 12 years. She has worked with leading corporate brands such
as Deloitte, NIIT and Ramoji Film City. She completed her Ph.D. from SLS, PDPU and
her area of study was the Digital India initiative by the Government of India.
Dr. Amrita Chakraborty, Identification of Public Sentiment over Comments
through Tweets by Digital India, -Palarch’s Journal Of Archaeology Of
Egypt/Egyptology 17(7),ISSN 1567-214x
Abstract: The Digital India initiative by the Government of India is an initiative by
Shri Narendra Modi, the Prime Minister of India. Launched in the year 2015, the
programme enhanced its scope to various digital services wings. In this study, the
researcher attempted to identify public sentiments by studying their comments through
Tweets. The researcher considered the Tweets by Digital India for a period of one year
from November 21, 2016 to November 20, 2017. With the help of Quasi Systematic
9661
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Sampling, 572 Tweets were pulled up for analysis. The researcher studied the first two
comments of each Tweet. Every comment was categorized based on primary sentiment
of the sentence such as ‘Support’, ‘Criticism’, ‘Neutral’, ‘Enquiry’ and ‘Out of
Context’. In order to validate the researcher’s conclusion of the sentiment on any
comment, the researcher conducted ‘Inter-Rater Reliability Test’ to calculate Kappa
Coefficient (k). This study intends to find out the sentiment of the public in majority
and the Twitter accounts, which received the higher percentage of supportive, neutral or
criticizing comments. The research further tries to identify if the supportive or
criticizing comments are higher in English or Hindi.
Keywords: Digital India, Sentiment Analysis, Twitter, Public Relations, Twitter
Comments
Funding: This research received no specific grant from any funding
agency in the public, commercial or not-for-profit sectors.
Declaration of Conflicting Interests: The Author declares that there
is no conflict of interest.
Introduction: What exactly is the Digital India programme? While
many scholars use the phrase with relative impunity, it is often
misunderstood and misinterpreted. Before delving into the discussion
any further, it is important that we understand the concept in its
totality. The Government of India launched the Digital India
programme so that essential government services are provided to the
citizens through electronic means. The principal mechanism to
achieve the same was a complete revamping of the online
infrastructure and digital empowerment of the people.
The official website of Digital India defines the programme thus,
“The Digital India programme is a flagship programme of the
Government of India with a vision to transform India into a digitally
empowered society and knowledge economy” (Digital India, 2019).
If we take a closer look at the definition, we shall appreciate that ‘eGovernance’ is the buzzword here. e-Governance by itself is a
relatively new concept in India.
9662
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
While the internet was launched in India back during the midnineties of the last century, it was largely an elite concept until the
turn of the millennium. The renowned Dot-com Bubble ensured that
India also started to gather momentum in terms of the number of
internet users. The trend has continued ever since and India today has
one of the largest internet subscriber bases globally. According to the
Internet and Mobile Association of India (IAMAI), roughly 35 per
cent of the Indian populace (close to 481 million people) uses the
World Wide Web as of December 2017 (Agarwal, 2018). This has
inspired multiple government authorities to adopt e-Governance,
which truly has the potential to transform the country into a
knowledge economy.
If we take a good and hard look at the history of e-Governance in
India, we shall find that the concept started taking shape right after
the introduction of internet in the country. Of course, the emphasis
back then was decidedly on citizen-centric services. Slowly but
gradually, a host of states and Union Territories initiated a number of
e-Governance schemes. As a case in point, Andhra Pradesh
successfully tested a number of e-Governance initiatives. The state’s
government in fact clinched three gold awards at the 20th National
Conference on e-Governance in 2017 (The Hindu, 2017).
It will be interesting to note that the purpose of e-Governance in
India has changed over the years. Initially, it revolved around the
computerization of records at the government departments. Now, the
principal objectives of e-Governance are citizen-centricity and
transparency. Notwithstanding all these developments, e-Governance
could not make much of headway in India. The villages were still to
be covered and the lower strata of the society were yet to be
encompassed. To address these gaps, the National e-Governance Plan
(NeGP) was launched. The idea was to take a complete view of all
the possible e-Governance schemes in the country and bring them
9663
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
under the same roof. The plan was initiated on May 18, 2006. It
initially had 27 Mission Mode Projects (MMPs) and eight
components. In 2011, four MMPs were added to take the tally to 31.
At the end of the day, the intent was to make public services more
accessible to the common people, as can be gathered from the Vision
Statement of NeGP (Aggarwal, 2018).
Despite numerous success stories, e-Governance could not still make
the impact that that it was supposed to. Successive governments have
felt the necessity of promoting inclusive growth through the wider
implementation of e-Governance. The most important areas that were
needed to cover were electronic services, products, devices and job
opportunities. It was also felt that extra emphasis was needed to be
put on electronic manufacturing in the country.
To address all these issues and many more, the Government of India
kick-started the Digital India programme. The programme is
expected to transform the country into a digitally empowered society
(Digital India, 2019).
The Digital India programme has three key vision areas:
1. Digital Infrastructure as a Core Utility to Every Citizen.
2. Governance and Services on Demand.
3. Digital Empowerment of Citizens (Digital India, 2019).
Digital India is an encompassing programme that covers a number of
ministries and departments under the Government of India (GoI).
Thus, the emphasis areas of the programme also range across many
parallel domains. To streamline the functioning of the programme,
the authorities have identified nine specific pillars. Each of the pillars
falls under multiple ministries and departments. The pillars are:
1. Broadband Highways.
2. Universal Access to Mobile Connectivity.
3. Public Internet Access Programme.
4. e-Governance – Reforming Government through Technology.
9664
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
5. e-Kranti – Electronic Delivery of Services.
6. Information for All.
7. Electronics Manufacturing.
8. IT for Jobs.
9. Early Harvest Programmes (Digital India, 2019).
Reasons for Studying Twitter: Twitter calls itself an information
network. It stands in stark contrast with platforms such as Facebook,
where information is incidental. Although there exist very few
statistical studies vis-à-vis the exact status of Twitter as a news
aggregator, the fact remains that a majority of the news organizations
make it a point to post all their breaking news items on Twitter. Even
a cursory glance through the Twitter page of a news organization
establishes the primordial nature of Twitter in the world of news. If
we follow the different social media handles of Digital India, we
shall be able to appreciate that the maximum number of posts
concerning the programme are published on the Twitter page itself.
In addition, in India, a number of political and social developments
are discussed over Twitter. Bureaucrats, politicians and public
figures regularly post theirs views and opinions on Twitter thereby
making it the dominant social medium.
Objectives of the Study: The study focusses on finding objective
answers to the following three problem questions:
1. What are the various categories of public comments and what are
their respective numbers?
2. What are the top Twitter accounts that received the highest
comments falling under Support, Neutral, Criticism and Enquiry
categories?
3. Which are the categories of communication (languages/ links) used
in the comments by the public and what sentiments do they reflect?
Review of Literature
9665
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
This chapter makes a concerted attempt at summarizing the various
studies and literature available on Twitter as an effective social
media to understand public sentiment.
Research on Twitter as a Social Media Platform: There is a
largely theoretical paper that makes a distinct attempt at
understanding the medium that is Twitter. Titled ‘Understanding
Twitter’ and penned in 2012, the paper by Fiona Maclean, Derek
Jones, Gail Carin-Levy and Heather Hunter defines and introduces
the concept of Twitter within the domain of occupational therapy
research and consequent education. According to the paper, Twitter
is an accessible, versatile and valuable instrument for the global
communication of ideas, thoughts and visions for the future. The
paper further says although Twitter poses some distinct challenges,
the nature of the medium offers a rather exciting prospect that should
and can be embraced. In very simple terms, the paper vouches for the
usage of Twitter in multiple ways. Twitter is cross-disciplinary in
nature and people from all walks of life can connect and network
with each other through the platform. This paper again establishes
the importance that Twitter has assumed over the past few years. The
current researcher has time and again asserted the reason as to why
Twitter has been selected over other social media platforms. This
paper confirms and validates that choice. As people are moving
ahead, different organizations are adopting Twitter in a big way.
As fundamental as it sounds, there is a paper on how Twitter is used
by different users. Titled ‘What Do We Do with Twitter?’ and
published in 2014, the paper penned by Jimmy Sanderson is a treatise
to find out the reason as to why papers on Twitter have multiplied
over the past few years. The paper also tries to find out as to why the
influence of social media in general and Twitter in particular is
increasing when it comes to sports. However, the author also
contended that while impact on Twitter is being measured currently,
9666
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
the impact of Instagram or similar other forums will be measured in
the near future. This paper establishes the importance of Twitter in
sports as well. All in all, Twitter has emerged as one of the most
important platforms of communication in practically all the fields.
The current researcher maintains that understanding Twitter has
become absolutely important in understanding the dynamics of
contemporary life. According to this paper, thematic analyses could
also be done using Twitter by analyzing the content available on the
platform. In very simple terms, what is being communicated through
Twitter about any of these fields also make a difference.
In an interesting commentary on Twitter, Dhiraj Murthy in 2011
brought out some rather stark facts. Effectively titled, ‘Twitter:
Microphone for the Masses?’, the commentary explores Twitter and
citizen journalism. As has been noted earlier, Twitter has been
responsible for breaking multiple stories at multiple points in time. In
fact, the commentary bats for the fact that Twitter has converted
normal users into citizen journalists. However, the commentary also
tries to find out if the traditional media subsumes the voices on
Twitter with its massive presence. The study finds out that subject to
the influence of Twitter, news organizations are increasingly
Tweeting the headlines of breaking news items, which again
underscores the importance of Twitter in journalism. Another issue
that comes up in the study is the fact that Twitter remains accessible
only to the technical literate thereby taking out the less privileged
ones out of the radar. Hoaxes and misinformation on Twitter have the
ability to create mess in the otherwise maintained news space.
However, there is one thing that comes out of the study – news
organizations the world over are increasingly changing the pattern
based on which they produce and disseminate news subject to the
infusion of Twitter in the scene.
9667
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Twitter’s Influence on Modifying Public Opinion: There is
another study that analyzes the power of Twitter in spreading
disinformation and fake news. Titled ‘Disinformation, ‘Fake News’
and Influence Campaigns on Twitter’, this study was conducted in
2018 by Matthew Hindman and Vlad Barash on behalf of the Knight
Foundation. With the background being the 2016 American
presidential election campaign, the study was one of the largest of its
kind that tried finding out the exact nature of fake news on Twitter.
The study also made an attempt at finding out how powerful
influence campaigns can be on Twitter. In very simple terms, the
study tried finding out the strength of Twitter in influencing
citizenry. Interestingly, the researchers were of the opinion that the
presence of multiple accounts on Twitter that spread disinformation
puts serious questions on the efforts made by Twitter to control fake
news. In fact, the primary finding of the study was that there are
multiple accounts that are often interlinked and that have a definite
agenda that maliciously plant fake information on Twitter to
misguide people and influence their actions. One can safely deduce
that Twitter has become probably the most important platform
through which public opinion can be manipulated and tampered with,
not to say that it can definitely be influenced.
There is another 2015 study that tries to figure out the suitability of
Twitter for demographic and social science research. Titled ‘Using
Twitter for Demographic and Social Science Research: Tools for
Data Collection and Processing’, the researchers Tyler H.
McCormick, Hedwig Lee, Nina Cesare, Ali Shojaie and Emma S.
Spiro contend that there is a significant potential for Twitter to be
used for social science research. This again establishes the
importance of Twitter in figuring out public opinion. However, the
biggest challenge lies in finding out the demographic details about
the users of Twitter. The study designs an appropriate data
9668
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
processing strategy for using Twitter in social science research. The
research also tries to find out the reliability of the techniques used.
One of the most important deductions of this research is the fact that
Twitter remains one of the best sources of information for conducting
any kind of social science research. This again underscores the
importance of the study in question. Using textual information from
Twitter can open up a new window of information. It can be argued
that the success or failure of any scheme can be gauged perfectly
with the help of an informational study through Twitter. The
researcher used Amazon’s Mechanical Truck (AMT) to code a large
volume of users profile images efficiently and made a comparison
between the evaluations of the truckers with expert coders. The
researcher used this method for data collection to conduct a racial
comparison. The researcher selected the truckers based on their
demographic details to justify the objective of the research.
There exists a myth busting study that establishes the exact nature of
Twitter research. The study conducted by Andrew Billings in 2014
and titled ‘Power in the Reverberation: Why Twitter Matters, But
Not the Way Most Believe’ explored the role of Twitter in setting the
agenda of the mainstream sports media. The study contends that
Twitter as a social media platform is worth analyzing not because it
represents the public in general but because it doesn’t represent the
population in general. Again, this study tries to find out the
importance of Twitter in another very important area of public
attention – sport. As one can understand, the preeminence of Twitter
in the overall scheme of public activity is established solidly with
this study as well. Interestingly, the followers of players and team are
quite high in sports. However, none of the top Twitter influencer is
from the arena of sports. The study also points towards the fact that
there are some silent followers who are on Twitter without Twitting.
This is a catchy area. One can’t really say anything about this section
9669
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
that silently follows Twitter activities. All in all, this study also
emphasizes on the importance of Twitter in measuring public
opinion.
There is another very interesting study that talks about the analysis of
Twitter information for evaluating public views on transportation
services. In the 2020 research paper titled ‘A framework with
efficient extraction and analysis of Twitter data for evaluating public
opinions on transportation services’, penned by M.E. Rinker, there is
an effort to explore public views on transportation services so that
effective policies and decisions can be taken. The paper puts forward
a comprehensive framework that lets the extraction and subsequent
analysis of opinions using Twitter. The framework is explained using
the transportation system of the Miami-Date County. The paper puts
forward the view that for any infrastructure project evaluation and
eventual planning, the role that public opinion plays is vital. This is
exactly where social media can come handy. It can provide an extra
stream of information that decision makers and managers can use for
comprehending the public perception concerning the transportation
services. This paper again proves the importance of Twitter in
understanding and mining public perception concerning any activity.
Whether it is the government or any other entity carrying out any
massive public-facing activity, it is compulsory to look at Twitter
and find out how the public perception is shaping up. This research
also provides a broad framework under which similar studies can be
conducted in the future.
Sentiment Analysis on Twitter: In a rather interesting study titled
‘Diabetes on Twitter: A Sentiment Analysis’ in 2018, Elia Gabarron,
Enrique Dorronzoro, Octavio Rivera-Romero and Rolf Wynn tried
analyzing the sentiments on Diabetes through the platform of
Twitter. They believed that contents published on social media in
general and Twitter in particular have a strong impact on public
9670
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
sentiments on any issue and diseases are no exception. The study
made an explicit attempt at deconstructing the sentiments embedded
with the messages on Diabetes on Twitter. In fact, the intent of the
study was to find out how public health strategies could be fixed
using content on Twitter. The researchers were of the opinion that the
public perception about the disease could be found out using the
messages that were posted on Twitter. Finding out the perception of
common people and people affected by diabetes could significantly
help fix the trajectory along which public strategies could be
designed to help streamline the situation. In this case, the researchers
made a distinction between positive sentiments, negative sentiments
and neutral sentiments. One can take the example of the use of
emojis in the messages. The use of emojis was considered to be
showing a positive sentiment. The study could go a long way in
developing a constructive attitude among people towards diabetes.
There is another 2018 study that analyzes the sentiments on Twitter
rather aptly. Titled ‘A Mixed-Methods Examination of Morality
Work through Sentiment Analysis and Qualitative Coding of Twitter
Data’ by Tony P. Love, Jenny L. Davis and Joseph M. Calvert, the
researchers analyze public sentiments over Twitter. One can assert
that public sentiments vary on Twitter depending on multiple
parameters. The study shows how public sentiments shift on Twitter
depending on the target of the moral judgment. The study details how
discourses on Twitter often point towards a dynamic nature of public
sentiments. The study outlines a case study and shows the differences
in public sentiments. It shows how social media reactions were
mostly negative when Ray Rice, an American football player,
punched his fiancée. However, when the fiancée married Ray Rice,
the sentiments on social media turned less negative although there
was a case of abuse earlier. The fickle nature of Twitter discourses
could be understood through this episode. It is clear that Twitter still
9671
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
needs some attention to understand the drift of how the members of
the public react to specific situations. Another point that was argued
by the researchers is that moral boundaries are drawn and redrawn
with each debate on Twitter.
In this context, one needs to talk about a very important study that
was conducted by Yinying Wang and David J. Fikis in 2017. Titled
‘Common Core State Standards on Twitter: Public Sentiment and
Opinion Leaders’, the researchers tried examining the public opinion
using the Common Core State Standards (CCSS) on Twitter in the
United States of America. Sentiment analysis was done as a part of
the research and users of Twitter expressed their discontent towards
CCSS in all the 50 states of the country. It was found out that most of
the opinion leaders were in the given issue were the ones who
expressed negative sentiments on the issue. This study is again a very
strong indication of how Twitter can be a great platform to
understand public sentiments over multiple issues. While this
research was centered in the United States of America, similar
studies on similar issues in India could throw up rather interesting
results. This researcher found out the relevance of all Twitter studies
before ascertaining its importance in the study in question. As one
can understand, Twitter could be a great mining platform through
which
the
popularity
of
various
governmental
and
semi-
governmental schemes could be enhanced significantly.
Research Design and Methodology: The researcher used Content
Analysis. Quasi Systematic Sampling (or Systematic Like Sampling)
was used for fulfilling the objective. The Sample Size was 572 out of
a Population of 2690.
The researcher used all the 572 Tweets for study out of which 170
Tweets have no comment. Hence, they were categorized as N/A (Not
Applicable). The researcher selected two comments per Tweet. In the
process, 881 comments by public were studied. The researcher found
9672
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
out that the order of comments by public in a Tweet changes from
time to time. Thus, the top two comments were used for the content
analysis.
The researcher conducted an Inter-Rater Reliability Test in
calculating Cohen’s Kappa (k) to find out the reliability or interobserver agreement between two raters A and B.
Inter-Rater
Reliability, or precision, happens when one’s data raters (or
collectors) give the same score to the same data item (Stephanie,
2014).
The researcher selected 30 comments from 881 comments of study
through Systematic Sampling by using the formula 881/30 = 29.36.
Hence, the researcher considered every 29th comment to study.
Considering the population size of 5262 comments by public with a
sample size of 881 comments for observation, with 95 per cent
confidence level, the margin of error comes to 3.01 per cent. At 99
per cent confidence level, the margin of error is 3.96 per cent.
The researcher categorized the comment category with the following
codes to study the Kappa Coefficient (k).
The researcher coded 1 for rater’s agreement with the researcher and
0 for rater’s disagreement with the researcher.
Results and Discussions: It is important to find out the answers to
each of the research questions in a systematic and objective manner:
1. What are the various categories of public comments and their
numbers?
This research question can also be divided into the following survey
questions, the answers to which will determine the composite
answer.
1.1. Category of Comments and Their Numbers.
9673
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Table 1
Category of Comments
Total
Percentage
Support
230
26
N/A
170
20
Neutral
135
15
Out of Context
134
15
Criticism
122
14
Enquiry
90
10
Grand Total
881
Figure 1
From the table and the figure above, the following pointers become
clear:
1) Supportive comments represent the highest category with a total
number of 230 comments or 26 per cent of total comments taken for
the study.
2) The next highest category of comments includes neutral
statements, which account for 135 comments or just 15 per cent of
the total comments.
9674
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
3) 15 per cent of the comments were out of context, which has no
relevance with the Tweets of Digital India.
4) There are 122 comments with criticism, which account for 14 per
cent of the comments being studied.
5) A number of individuals came up with enquiries for various
services of Digital India, which account for 90 comments or 10 per
cent.
6) The above pieces of data clearly indicate that the Twitter page of
the Digital India campaign significantly turned people’s opinions in
its favour. Although 26 per cent cannot be identified as
overwhelming public opinion, there are substantial reasons to believe
that the situation will only improve with each passing day. The fact
that internet pages would always invite irrelevant and out-of-context
comments cannot be ignored either. Probably, this is one of the
reasons as to why the support percentage is not as high as it should
have been.
2. What are the top Twitter accounts that received the highest
comments falling under Support, Neutral, Criticism and Enquiry
categories?
This research question can also be divided into the following survey
questions, the answers to which will determine the composite
answer.
2.1. Top Six Twitter Accounts with Highest supportive
Comments.
Table 2
Tweets Posted By
Count of
Total
Percentage
Supportive
Tweets
of
Comments Studied Acceptance
Digital India
127
541
23.48
Ravi Shankar Prasad
47
128
36.72
Ajay Kumar
7
12
58.33
9675
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
MyGovIndia
5
8
62.50
Ajay Sawhney
4
11
36.36
Grameen Vidyutikaran
4
5
80.00
Figure 2
From the table and the figure above, the following pointers become
clear:
1) With 36.8 per cent of supportive comments, it is evident that the
Tweets by Ravi Shankar Prasad were more accepted by the people.
2) Tweets by Digital India that were 541 in number were supported
only 127 times or with 23.47 per cent of supportive comments.
3) Ravi Shankar Prasad, who was the concerned Union Minister at
the time of the study, naturally attracted a lot of public attention and
acceptability. Here again, the support percentage, although
constituting the largest category, does not cross the halfway mark.
The reasons are fairly comprehensible and clear. There would always
be a dedicated section of the country’s populace, who would be
politically opposed to a certain minister. This section would continue
to spew their vitriol irrespective of the ground realities. The same
9676
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
could be applied to the case of Prasad’s Tweets concerning Digital
India initiative as well.
2.2. Top Five Twitter Accounts with Highest Neutral Comments.
Table 3
Tweets Posted By
Count of
Total
Percentage
Neutral
Tweets
of
Comments Studied Acceptance
Digital India
83
541
15.34
Ravi Shankar Prasad
27
128
21.09
NITI Aayog
5
27
18.52
NIELIT
4
10
40.00
Ajay Sawhney
3
11
27.27
Figure 3
From the table and the figure above, the following pointers become
clear:
1) The table shows that 21.09 per cent of comments through the
Tweets of Mr. Ravi Shankar Prasad are neutral.
9677
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
2) Prasad received the higher percentage of neutral comments as
compared to the comments received by Digital India account, which
is 15.34 per cent.
3) Neutral comments neither express support nor express
disapproval. Here as well, the results were not out of expectations.
Prasad, who was the concerned minister in charge at the time of the
study, naturally is the person to get in touch with. Naturally,
therefore, the comments that belong to the neutral category flooded
his Tweets. At the same time, the Twitter page of Digital India was
not left high and dry in the process. This page also had its share of
neutral comments, comments that could become positive if there is a
substantial push in the publicity activities of the page.
2.3. Top Five Twitter Accounts with Highest Comments on
Criticism.
Table 4
Tweets Posted By
Count of
Total
Percentage
Criticizing
Tweets
of
Comments Studied Acceptance
Digital India
76
541
14.05
Ravi Shankar Prasad
25
128
19.53
NITI Aayog
4
27
14.81
Aadhaar
3
4
75.00
Ajay Kumar
3
12
25.00
9678
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Figure 4
From the table and the figure above, the following pointers become
clear:
1) The above table shows that with a lesser number of Tweets,
people criticized the Tweets of Ravi Shankar Prasad more than
Digital India Tweets by 5.03 per cent.
2) It needs to be remembered here that the chances of netizens
criticizing a person is more than the chances of the same netizens
criticizing an official page. Subject to the fact that Prasad has a
human face, he was at the receiving end of most of the criticizing and
abusing comments. Not all abusive and criticizing comments could
be attributed to the performance factor associated with Digital India.
Some of the comments are made just for the heck of making
comments. The Digital India Tweets by Prasad also invited some
unwanted and ill-informed criticizing comments from netizens across
the spectrum.
2.4. Top Five Twitter accounts with highest comments on
enquiry.
9679
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Table 5
Count of
Comments
Total
Percentage
with
Tweets
of
Tweets Posted By
Enquiry
Studied Acceptance
Digital India
55
541
10.17
Ravi Shankar Prasad
13
128
10.16
GCCS Official
5
44
11.36
NITI Aayog
4
27
14.81
Ajay Sawhney
3
11
27.27
Figure 5
From the table and the figure above, the following pointers become
clear:
1) The percentage of enquiries addressed to Digital India and Ravi
Shankar Prasad are almost the same. Ajay Sawhney received three
enquiries from 11 posts. Those enquiries were mainly on National
Digital Payments Mission, the extent of digital platform in the
Parliament and its adoption by the Member of Parliament. One of the
comments was also for seeking more information on the TCS
9680
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
initiatives to on-board and train merchants on the Aadhaar-enabled
platform.
2) The Digital India initiative is an innovative programme launched
by the Government of India. No wonder that there are queries vis-àvis its implementation, scopes and plans. Therefore, many netizens
wanted to know as to what the scheme is all about. Subject to the fact
that the programme has nine different pillars, the enquiries ranged
across multiple domains. While some of the enquiries were genuine,
some were evidently spurious in nature. However, the nature of the
internet is such that there would always be a bunch of people, who
would ask the wrong questions and for the heck of asking.
3. Which are the categories of communication (language/ link)
used in the comments by public and what sentiments do they
reflect?
This research question can also be divided into the following survey
questions, the answers to which will determine the composite
answer.
3.1. Languages of Comments.
Table 6
Total No.
of
Languages or Categories
Comments
English
608
Hindi
99
Link
5
Image
3
Punctuation
1
None Noted
165
9681
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Figure 6
From the table and the figure above, the following pointers become
clear:
1) Out of 881 comments studied in this research, the researcher found
that 204 comments or 69 per cent of the total comments are in the
English language.
2) There were 165 Tweets, which received no comments from
people. Those Tweets were from various accounts and covered all
the eight pillars of the Digital India programme.
3) There were 99 comments, which were posted in the Hindi
language. They account for 11 per cent of the total posts.
4) The findings that were arrived above cement some convictions
about the state of internet in the country. It remains a fact and is
proven beyond doubt by the first finding that English is the dominant
language of communication followed by Hindi, which is the most
spoken language in the country. If one takes a closer look, it can be
seen that English establishes its dominance over Hindi by a distance.
There were quite a few Tweets that did not attract any comment and
could not be categorized based on languages. It is also interesting
that very few Tweets attracted links as comments.
9682
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
3.2. Break-Up of the Languages of Comments.
Total
Context
Out of
Neutral
Comments
No
Enquiry
Language
Criticism
Supportive
Table 7
English
204
103
84
0
121
96
608
Hindi
23
19
5
0
14
38
99
Link
0
0
0
5
0
0
5
Image
3
0
0
0
0
0
3
Punctuation
0
0
1
0
0
0
1
None Noted
0
0
0
165
0
0
165
Grand Total
230
122
90
172
135
132
881
Figure 7
From the table and the figure above, the following pointers become
clear:
1) The highest number of supportive coments were in the English
language.
9683
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
2) Out of the 608 comments in English, the percentage of supportive
comments account for 33.55 per cent, criticising comments account
for 16.94 per cent and neutral comments account for 19.90 per cent.
3) Out of the 99 comments in Hindi, supportive comments account
for 23.23 per cent, criticising comments account for 19.19 per cent
and neutral comments account for 14.14 per cent.
4) Thus, the percentage of supportive and neutral comments are
higher in English language by 10.32 per cent and 5.76 per cent
respectively. The percentage of criticising comments are higher in
the Hindi language by 2.25 per cent.
5) It is quite interesting to note that the percentage of supportive
comments in English is significantly higher than the percentage of
supportive comments in Hindi. To an extent, it establishes the citycentric orientation of any digital initiative, with Digital India
programme not constituting an exception. The same thing happens
when one looks at the neutral comments as well. Here as well,
English trumps Hindi in terms of sheer percentages. In very simple
words, it could be deduced that the acceptability factor of the Digital
India initiative is much higher amongst the English educated class.
Inter-Rater Reliability Test to Calculate Cohen’s Kappa (k) for
Objective 3: The researcher conducted Inter-Rater reliability test to
calculate Cohen’s Kappa (k) to find out the reliability or
interobserver agreement between two raters A and B.
Interrater
reliability, or precision, happens when one’s data raters (or
collectors) give the same score to the same data item.
The researcher selected 30 comments from 881 comments of study
through Systematic Sampling by using the formula 881/30 = 29.36.
Hence, the researcher considered every 29th comment to study.
The researher categorized the ‘comment category’ with the following
codes to study the Kappa Coefficient (k)
9684
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
The researcher further coded 1 for rater’s agreement with the
researcher and 0 for rater’s disagreement with the researcher.
Table 8
Category
Code
Support
A
Neutral
B
Criticism
C
Enquiry
D
Out of Context
E
N/A
F
Table 9
Agreement
1
Disagreement
0
Formula for Calculating Cohen’s Kappa:
Here, po is the relative observed agreement among raters and pe is the
hypothetical probability of chance agreement.
The Kappa Coefficient (k) varies from 0 to 1, where:
0 = Agreement Equivalent to Chance.
0.1 – 0.20 = Slight Agreement.
0.21 – 0.40 = Fair Agreement.
0.41 – 0.60 = Moderate Agreement.
0.61 – 0.80 = Substantial Agreement.
0.81 – 0.99 = Near Perfect Agreement
1 = Perfect Agreement.
Data Observation:
9685
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
1) 27 conclusions by researcher were accepted by both Rater A and
B.
2) 2 conclusions by researcher was rejected by both Rater A and B.
3) Overall Rater A accepted 28 conclusions and rejected 2
conclusions by the researcher.
4) Overall Rater B accepted 27 conclusions and rejected 3
conclusions by the researcher.
Calculation of Cohen’s Kappa for This Data Set:
1) Calculation of po, the relative observed agreement among raters:
27 conclusions by the researcher were accepted by both the Raters A
and B.
2 conclusions by the researcher was rejected by both the Raters A
and B.
Therefore, po = Number in agreement/Total = (27 + 2)/30 = 29/30 =
0.96
2) Calculation of pe, the hypothetical probability chance agreement:
Probability of both the raters agreeing to researcher:
Rater A accepted 28 conclusions out of 30, i.e. 28/30 = 0.93.
Rater B accepted 27 conclusions out of 30, i.e. 27/30 = 0.9.
Total probability of both the raters accepting the researcher = 0.93 x
0.9 = 0.84.
Probability of both the raters rejecting to the researcher:
Rater A rejected 2 conclusions by the researcher out of 30, i.e. 2/30 =
0.06.
Rater B rejected 3 conclusions by the researcher out of 30, i.e. 3/30 =
0.1.
Total probability of both the raters rejecting the researcher is 0.13 x
0.1 = 0.006.
Therefore, pe = 0.84 + 0.006 = 0.85.
Therefore, applying the values of po and pe in the formula of Cohen’s
kappa,
9686
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
= (0.96 - 0.85)/(1 - 0.85)
= 0.11/0.15
= 0.73
This final value falls in the range of 0.61 – 0.80. Hence, there is
substantial agreement between the two raters.
Important Notes:
In context of the study done to fulfill Objective 3, it is pertinent that
some of the words are explained.
1. Support represents positive comment. In other words, support is
categorized as any affirmative comment given in favour of any of the
schemes of the Digital India initiative.
2. Criticism represents negative comment. Criticism could be defined as
any negative comment given concerning any of the schemes of the
Digital India initiative.
3. Neutral statements represent statements related to the post, but not
for or against the programme. No content analysis on neutral
statements has been done purely because the same cannot be
segregated any further.
4. Enquiry means any enquiry on the Digital India initiative. No content
analysis on enquiries has been done again because the same cannot
be segregated any further.
5. Comments are marked as English or Hindi as per the script used for
making the comment.
6. News coverage as comment is studied on the basis of the headline.
7. Full sentence or combination of words is considered for content
analysis in order to understand the context of support or criticism.
8. If comments are mixed in Hindi and English script, then the script for
the first word has been considered for study.
9. Comments with image have not been considered for study.
9687
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Consecutive comments by same person is not taken for content
analysis for the second time. Consecutive comments by same person
is taken for content analysis in cases where there are only two
comments.
Conclusion: To summarize the analysis, it can be primarily said that
the public comments on the Digital India Twitter page could be
broadly categorized into supportive, neutral, criticism and enquiry
categories. With the numerical calculation involved in the content
analysis, it was found out that supportive comments represent the
highest category with a total number of 230 comments or 26 per cent
of total comments (881) taken for the study. The next highest number
of comments fell under the neutral category, which accounted for
135 comments or just 15 per cent of the total comments. There were
122 comments with criticism, which accounted for 14 per cent of the
comments being studied. A number of individuals came up with
enquiries for various services of Digital India, which accounted for
90 comments or 10 per cent. 15 per cent of the comments were out of
context – comments that had no relevance with reference to the
Digital India tweets.
With 36.8 per cent of supportive comments, it is evident that the
Tweets by Ravi Shankar Prasad were more accepted by the people.
Tweets by Digital India, with the number being 541, were supported
only 127 times translating to 23.47 per cent. 21.09 per cent of
comments through the Tweets of Ravi Shankar Prasad were neutral.
Prasad got the higher percentage of neutral comments as compared to
the comments received by Digital India account, which stood at
15.34 per cent. With a smaller number of tweets, people criticized the
Tweets of Ravi Shankar Prasad more than Digital India Tweets by
5.03 per cent. The percentage of enquiries asked to Digital India and
Ravi Shankar Prasad were almost the same. The primary reason for
9688
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
Ravi Shankar Prasad having received the maximum number of
comments in all categories is his significantly large number of
Twitter followers as compared to the Digital India account.
Out of the 881 comments studied in this research, the researcher
found that 204 comments or 69 per cent of the comments were in the
English language. There were 165 Tweets, which received no
comments from people. There were 99 comments or 11 per cent of
the total comments, which were posted in the Hindi language. Hence,
a significantly high number of comments on the Twitter handle of
Digital India were in English.
The analysis further highlighted that the highest number of
supportive comments was in the English language. Out of the 608
comments in English, the percentage of supportive comments stood
at 33.55, neutral comments stood at 19.90 per cent and criticizing
comments stood at 16.94 per cent. Out of the 99 comments in Hindi,
supportive comments stood at 23.23 per cent, neutral comments
stood at 14.14 per cent and criticizing comments stood at 19.19 per
cent. Thus, the percentages of supportive and neutral comments were
higher in English language by 10.32 per cent and 5.76 per cent
respectively. The percentage of criticizing comments was higher in
the Hindi language by 2.25 per cent. Hence, public comments that
were posted in English had a higher inclination towards supporting
the Tweets by Digital India.
Limitations: Although the researcher tried to make the study as
error-free as possible, there were some limitations that crept into the
study. One can look at some of the bigger limitations first. Here they
are:
1. The total number of Tweets posted on the Digital India Twitter page
was 2,690 for a period of one year ranging from 21/11/2016 to
9689
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
20/11/2017. Due to limitation on time available for the study, the
researcher could not consider all the Tweets for study and hence
obtained the data with Quasi-systematic Sampling (or Systematiclike Sampling) by taking a representation of two days per week,
maintaining a gap of seven days between each set.
2. The total number of comments for 572 Tweets was 5,262. Due to
limitation of time, the researcher studied only the first and second
comments of each Tweet, which translates to 881 comments.
3. Only the first tweet for each of the days was considered for the study.
4. The total number of replies in 572 posts was 5262. Therefore, the
average number of replies per Tweet is 9.199, which is equivalent to
10 posts per day. Hence, the researcher decided to consider the first
comment of each day due to the limitation of time.
5. The final comment is ranked as either Support, Neutral, Enquiry,
Criticism or Out of Context.
6. Some comments were given only with YouTube links. Those links
that do not play were not considered for the study.
7. The categorization of comments written in both English and Hindi
was done based on the script with which the comments start.
To sum it up, it could be safely deduced that the Twitter page of the
Digital India initiative has been significantly successful in making
people aware about the initiative turning their opinion in favour of
the programme. While there is no denying that the programme itself
has the potential to transform the fortunes of the country, the public
relations activities of its official Twitter page cannot be discounted
either. At a time when branding something properly has become the
cornerstone of success for any initiative, the Twitter page of Digital
India can be considered a good example of how such a page should
function.
References:
9690
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
1. Aggarwal, S. (2018, June 20). National e-Governance Plan.
Retrieved
from
National
e-Governance
Plan
Web
Site:
https://nisg.org/files/documents/A05140001.pdf
2. Digital India, M. o. (2019, March 13). About Digital India. Retrieved
from
Digital
India
Website:
http://www.digitalindia.gov.in/content/about-programme
3. Digital India, M. o. (2019, March 13). Digital India Power to
Empower.
Retrieved
from
Digital
India
Web
Site:
https://www.digitalindia.gov.in/
4. Digital India, M. o. (2019, March 13). E-GOVERNANCE POLICY
INITIATIVES UNDER DIGITAL INDIA. Retrieved from Digital
India
Web
site:
http://digitalindia.gov.in/content/e-governance-
policy-initiatives-under-digital-india
5. Digital
India,
M.
o.
(2019,
March
13).
ELECTRONICS
MANUFACTURING. Retrieved from Digital India Web site:
http://digitalindia.gov.in/content/electronics-manufacturing
6. Digital India, M. o. (2019, March 13). How Digital India will be
realized: Pillars of Digital India. Retrieved from Digital India
Website: http://digitalindia.gov.in/content/programme-pillars
7. Digital India, M. o. (2019, June 19). Vision of Digital India.
Retrieved
from
Digital
India
Website:
https://www.digitalindia.gov.in/content/vision-and-vision-areas
8. Reuters
(2017,
October
2).
Smartphones
made
in
India?
Manufacturing ambition hits hurdles. Retrieved from ET tech:
https://tech.economictimes.indiatimes.com/news/mobile/smartphones
-made-in-india-manufacturing-ambition-hits-hurdles/60911110
9. Government of India. (2018, June 19). Digital Payment Methods.
Retrieved
from
Cashless
India
Web
site:
http://cashlessindia.gov.in/digital_payment_methods.html
9691
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
10. PMINDIA. (2018, June 19). Make In India. Retrieved from PM India
Web site: http://www.pmindia.gov.in/en/major_initiatives/make-inindia/
11. Press Information Bureau (2015, March 25). Approach and Key
Components of e-Kranti: National e-Governance Plan 2.0. Retrieved
from
Press
Information
Bureau
Website:
http://pib.nic.in/newsite/PrintRelease.aspx?relid=117690
12. Press Information Bureau (2017, March 3). India to Host 10TH
International Conference on Theory and Practice of Electronic
Governance: ICEGOV 2017. Retrieved from Government of India
Web site: http://pib.nic.in/newsite/PrintRelease.aspx?relid=158823
13. Press Information Bureau (2017, July 21). India to Host Global
Conference on Cyber Space 2017 – World’s Largest Conference on
Cyber Space. Retrieved from Government of India Web site:
http://pib.nic.in/newsite/PrintRelease.aspx?relid=168850
14. Statista (2019, April 27). Penetration of leading social networks in
India as of 3rd quarter 2017. Retrieved from The Statistical Portal
Web
site:
https://www.statista.com/statistics/284436/india-social-
network-penetration/
15. Agarwal S. (2018, February 20). Internet users in India expected to
reach 500 million by June: IAMAI . Retrieved from The Economic
Times
Website:
https://economictimes.indiatimes.com/tech/internet/internet-users-inindia-expected-to-reach-500-million-by-juneiamai/articleshow/63000198.cms
16. FE Online (2018, May 31). BHIM app receiving this big update:
Know what it is and how to use it. Retrieved from Financial Express
Web
site:
https://www.financialexpress.com/industry/technology/bhim-appreceiving-this-big-update-know-what-it-is-and-how-to-useit/1188686/
9692
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
17. Fiona Maclean, D. J.-L. (2013). Understanding Twitter. British
Journal of Occupational Therapy, 295-298.
18. Sanderson,
J.
(2014).
What
Do
We
Do
With
Twitter?
Communication & Sport, 127-131.
19. Murthy, D. (2011). Twitter: Microphone for the masses? Media,
Culture & Society, 779-789.
20. Matthew Hindman, V. B. (2018). Disinformation, ‘Fake News’ and
Influence Campaigns on Twitter. Miami: Knight Foundation.
21. Tyler H. McCormick, H. L. (2017). Using Twitter for Demographic
and Social Science Research: Tools for Data Collection and
Processing. Sociological Methods & Research, 390-421.
22. DWIVEDI, VEDVYAS JAYPRAKASH, and Yogesh C. Joshi.
"PRODUCTIVITY IN 21 st CENTURY INDIAN HIGHER
EDUCATION INSTITUTIONS." International Journal of Human
Resource Management and Research 9.4 (2019): 61-80.
23. Billings, A. (2014). Power in the Reverberation: Why Twitter
Matters, But Not the Way Most Believe. Communication & Sport,
107-112.
24. Rinker, M. (2020). A framework with efficient extraction and
analysis of Twitter data for evaluating public opinions on
transportation services. Travel Behaviour and Society, 10-23.
25. Radhika, R., and Ramesh Kumar Satuluri. "A study on life insurance
penetration in India." International Journal of Human Resource
Management and Research (IJHRMR) 9.1 (2019).
26. Elia Gabarron, E. D.-R. (2018). Diabetes on Twitter: A Sentiment
Analysis. Journal of Diabetes Science and Technology, 1-6.
27. Tony P. Love, J. L. (2018). A Mixed-Methods Examination of
Morality Work Through Sentiment Analysis and Qualitative Coding
of Twitter Data. SAGE Research Methods Cases, 1-12.
28. JINDAL, NEENA, KRITIKA THAKUR, and TANIA SHARMA.
"DIGITAL INDIA: CHALLENGES, SOLUTIONS AND ITS
IMPACT ON SOCIETY." International Journal of Environment,
Ecology, Family and Urban Studies (IJEEFUS)9. 2, Apr 2019, 83-90
9693
Identification of Public Sentiment over Comments through Tweets by Digital India
PJAEE, 17 (7) (2020)
29. Yinying Wang, D. J. (2017). Common Core State Standards on
Twitter: Public Sentiment and Opinion Leaders. Educational Policy,
1-34.
30. Vanan, C. KURINCHI, and R. Subramani. "Digital divide: rural and
urban college students ‘attitude towards technology
acceptance." International Journal of Communication and Media
Studies (IJCMS) 5.4 (2015): 1-8.
9694