Sentiment Analysis for Election Campaigns

This article discusses how sentiment analysis can give a competitive edge to election campaigns by analyzing the sentiments of tweets related to the candidates. It explores the use of sentiment analysis to support a candidate's campaign and undermine the opponent's campaign. The article also highlights the role of data and insights mining in identifying trends and characteristics that may escape the attention of human analysts. Additionally, it addresses the ethical considerations involved in the development of tweet-bots for candidates.

Report

TABLE OF CONTENTS
Activity 2: Individual Case..............................................................................................................3
Q1: Sentiment Analysis...............................................................................................................3
Q2: Data / Insights Mining..........................................................................................................5
Q3: Ethics....................................................................................................................................6
REFERENCES................................................................................................................................8

Activity 2: Individual Case
Q1: Sentiment Analysis
Using examples from the articles above, and the sentiment analysis lectures, discuss how
sentiment analysis can give your candidate’s election campaign a competitive edge.
Discuss how sentiment analysis can be used to support your candidate’s campaign and
undermine your opponent’s campaign.
Place your response in the space below:
This can be observed from the articles that sentiments can be analysed through the tweets which
were made at the time of US 2020 Presidential election. Social media always play a major role in
such situations of race towards White House as this became the most powerful tool for the
individuals in communicating unconditionally. Twitter especially was used for analysing and
forecasting stock markets, international relations, disease outbreaks and elections. So the
sentiment analysis focussed on the opinions of the people on Twitter which changes towards the
candidates. In this the tweets are collected in which @realDonaldTrump and @JoeBiden are
mentioned. Then, the sentiments are analysed and the positivity/negativity ratio are compared.
This can be evaluated that Donald Trump has more followers with 86.3M followers while the
number of followers of Joe Biden were only 9.8M. Furthermore, this can also be examined that
17.7% of the tweets are positive for Biden while 13.7% for Trump. On the other hand, 39%
tweets were negative for Biden while the other 48% were negative for Trump (Xia, Yue and Liu,
2021). This interprets that the people were more positive towards Biden as compared to Trump.
This can also be evaluated that these views kept changing and the Trump sometimes got positive
tweet than Biden. This reaction was considered to be remarkable (Nugroho, 2021). But overall
the sentiments of the people towards Biden were more positive which can be seen through the
tweets on Twitter.
3

4

Activity 1: Group Case
Q2
Place your response and charts in the space below:
Handle Positive Negative Ratio-july 2016
Trump 519 419 1.24
Hillary 246 112 2.2
False, in July the PN ratio was not similar in both candidates. In Hillary ratio was 2.2 and in
trump it was 1.2. sThis is because there was high use of twitter in campaign. Besides that, there
was tough competition in between both candidates. Moreover, there was a great impact on
sentiments of tweet that was being done.
Row Count of ratio jan
5

Labels time
Negative 34 3
Positive 116 7
Grand
Total 150
Row
Labels
Count of
time ratio feb
Negative 23 1
Positive 69 3
Grand
Total 92
Row
Labels
Count of
time ratio march
Negative 37 4
Positive 51 6
Grand
Total 88
Row
Labels
Count of
time ratio april
Negative 18 3
Positive 52 4
Grand
Total 70
Row
Labels
Count of
time ratio may
Negative 19 3
Positive 58 4
Grand
Total 77
Row
Labels
Count of
time ratio june
Negative 39 4
Positive 55 5
Grand
Total 94
6

Row
Labels
Count of
time ratio july
Negative 36 4
Positive 78 6
Grand
Total 114
Row
Labels
Count of
time ratio Aug
Negative 57 4
Positive 77 6
Grand
Total 134
Q2: Data / Insights Mining
Using examples from the articles above, and the insights mining lectures, discuss how data and
insights mining can give your candidate’s election campaign a competitive edge.
Discuss how data and insights mining can help to identify trends and characteristics that may
escape the attention and awareness of human analysts.
Place your response in the space below:
Data is considered to be a gold mine of insight. This can help in increasing the unlock hidden
profitability, reducing client churn and also increasing the customer loyalty. Both the candidates
in the political parties realise the impact of predictive analytics and Big Data on the election
results. This can be observed that the tweet volume was considered as the biggest predictive
variable which helped in examining the popularity of the individual among the population. This
was observed that sentiments could change because of the campaigns and messages conveyed by
the parties but the data and insights mining. The data does not involve sarcasm and which is why
it does not skew the results. This can give a clear outline to the parties about the ways in which
they must implement their practices and strategies. The data also play a significant role in
7

electoral success. The sentiments can change according to the engagement and attention matrix
but the attention does not change the data and the insights mining (Wu, 2021). The attention does
not lead to change in the result because it may be possible that Trump got more attention due to
negative things which does not lead to increase in the votes. Therefore, data and insights mining
are the most appropriate method to measure the competition and helps in gaining competitive
advantage.
7.5 marks
8

Q3: Ethics
Your firm has been engaged to develop a tweet-bot for your client - either Hillary or Trump. The
tweet bot will data mine tweets associated with your candidate, their opponent and everyone
else who engages with these people in the form of likes, mentions, retweets and replies.
The tweet bot will then use natural language processing and latent semantic analysis to learn the
way your candidate ‘speaks’. Once trained, the tweet bot will tweet on your client’s behalf
so that it sounds like the tweet came from the candidate.
Discuss the ethical considerations involved in the development of this bot. Your response need to
address the Laws of Robotics outlined in the Appendix (see last page)
Place your response in the space below:
Though Tweet bots helps in giving the replies on the behalf of clients after training them but
there is a need to consider various ethics by focussing on Law of Robotics because the client will
not actually reply the people but a sort of robot will be developed to do the same. This include
that the tweet-bot or the robot must not injure human being in any way through any type of
inaction which can lead to their harm. Also, this can be observed that the ethics considered while
developing the tweet-bot must include that they must follow all the orders given by the humans
but this must not contradict the first rule which is not harm or injury to humans. Along with
considering the safety of the humans, the robot must also be able to protect itself along with
adhering to both the above laws (Stieglitz and et.al., 2017). These laws of robotics must be
considered as ethical considerations while development of the robots or tweet-bot. Along with all
these laws, there is third plus one law which aims at protecting the humanity in any case of
development. This suggests that the humanity must be protected along with the protection of
humans as well as robot. The bots must not abuse humans and also must not accept abuse from
humans (Cresci and et.al., 2019). This is how, ethics must be considered while developing the
9

tweet bots for interaction and replying on behalf of client.
10