Group Report: Climate Change
Through The Eyes Of Social Media
DIGIB201: 19B
By Group C:
Info Removed
1|Page
Contents
Abstract.................................................................................................................................................. 3
Introduction ........................................................................................................................................... 3
Research Questions ............................................................................................................................... 3
Related Work ........................................................................................................................................ 4
Tracking Climate Change Opinions.................................................................................................... 4
Global Warming and public concern .................................................................................................. 4
Global awareness ................................................................................................................................ 4
Denial .................................................................................................................................................. 4
Methodology .......................................................................................................................................... 4
Google Trends..................................................................................................................................... 4
Netlytic................................................................................................................................................ 5
Gephi ................................................................................................................................................... 5
Discussion .............................................................................................................................................. 5
Google Trends: “Climate Change: ...................................................................................................... 5
Interest Over Time: ......................................................................................................................... 5
Interest By Region: ......................................................................................................................... 6
Related Topics and Related Queries: .............................................................................................. 6
Text Analysis: ..................................................................................................................................... 6
Word cloud: .................................................................................................................................... 7
Sentiment Analysis: ............................................................................................................................ 8
Trump:............................................................................................................................................. 8
Thunburg:........................................................................................................................................ 9
“Climate Change” Keywords:....................................................................................................... 10
Gephi Network Visualisations: ......................................................................................................... 11
Thunberg: ...................................................................................................................................... 11
Trump:........................................................................................................................................... 13
Thunberg, Trump and “climate change” keywords: ..................................................................... 14
Conclusion ........................................................................................................................................... 15
References ............................................................................................................................................ 16
2|Page
Abstract
This report aims to answer three specific research questions related to climate change and the
differing thoughts people have on the issue. The questions are; how do people’s ideas and
thoughts differ on the issue of climate change, what impact is created towards the issue of
climate change from people’s differing views and how has interest grown globally surrounding
the uprising issue of climate change.
We implemented the tools Google Trends, Netlytics and Gephi to answer these questions.
These tools allowed us to discover that it is highly common for people to carry different
opinions on climate change. We conducted research and found related work which gave us
inspiration into our investigation. The interest over time through Google search has
dramatically increased from October 2018 to present, although climate change has always been
a global epidemic, concerns have been extremely raised in the past year. Hence, to combat
climate change, prominent leaders in our society must speak out and attempt to make a change,
the results from our text analysis shows that this is not the case. Leaders such as Donald Trump,
are careless towards climate change and the crumbling environment whilst, young individuals
such as Greta Thunberg, are taking lead by voicing the drastic but real-life environmental
issues.
Introduction
There are various problems affecting our planet and future - war, terrorism, poverty. However,
one problem trumps them all which is , climate change. The man-made disaster is currently at
the forefront of global debates, with famous climate change deniers like Donald Trump
branding it a hoax and young activists like Greta Thunberg voicing the issue in front of the
United Nations. Climate change is affecting everyone, as rising temperatures continue to melt
the polar ice caps, rising sea levels threaten to envelop low lying islands, towns and cities
around the world. The heightened temperature is causing increased forest fires that are
destroying ecosystems and are threatening to wipe out entire species.
This is an incredibly important topic as people are unaware of how close we are to destroying
our planet, and the little time we have left to fix it. With the assistance of social media, we will
try to understand how people feel about climate change, how it affects them, and what is being
done to reverse its effects.
Research Questions
For our investigation, we will be analysing the environmental issue known as climate change.
We will investigate and analyse this issue using social media data collected through various
tools and the knowledge we have gained through practise. In order to assess the issue of climate
change, we will be posing the following research questions:
1) How do people’s ideas and thoughts differ on the issue of climate change?
2) What impact (positive or negative) is created towards the issue of climate change
from people’s differing views?
3) How has interest grown globally surrounding the uprising issue of climate change?
What triggered this awareness?
3|Page
Related Work
Tracking Climate Change Opinions
Research was carried out by An, Ganguly, Fang, Scyphers, Hunter & Dy (2014) using data
mining and machine learning to discover public sentiment on climate change based on millions
of tweets. This study demonstrated that using social media data to identify opinions on climate
change and major climate events is an inexpensive way to extract valuable findings. This
supports our project which will also rely on Twitter data. They detected an increase in negative
sentiment from events such as President Barack Obama’s order on climate change whereas
they found a decrease in negative sentiment about climate change when extreme weather events
occurred such as the Typhoon Haiyan. Their research concludes that Twitter users’ opinions
on climate change can change over time and after major climate events.
Global Warming and public concern
A recent study by Bergquist & Warshaw (2019) examines the change in climate concern from
1999 to 2017 in the US. Using extensive survey data, they researched whether there has been
a change in public opinion on climate change due to an increase in global temperatures (global
warming). The research found that at a national and state level, there was an association with
changes in temperature and public concern, although the effects were small. They concluded
that global warming is unlikely to make a significant shift in Americans perspective on the
importance of climate change.
Global awareness
Lee, Markowitz, Howe, Ko and Leiserowitz (2015) conducted a study aimed at discovering the
predictors of global public climate change awareness and risk perception. They found that the
highest levels of climate change awareness (over 90%) were in developed parts of the world
such as Europe, Japan and North America where understanding the anthropogenic cause of
climate change is the strongest predictor of risk perceptions. On the contrary, developing
countries such as Egypt, India and Bangladesh had over 65% of respondents claim they had
never heard of climate change and perception of local temperature change was the strongest
predictor.
Denial
Many people deny climate change as real or an issue needing human intervention. A study
conducted by McCright & Dunlap (2011) examined the denial amongst conservative white
males in comparison to other adults in the US. Their study argues that there is a conservative
white male effect on climate change denialism, due to this group defending the social and
economic system and defending other members who challenge climate change. This is
relevant to our research as we will be looking into Trump who can be categorised as a
conservative white male who denies climate change.
Methodology
In terms of answering the research questions, we will implement the following tools:
Google Trends
Google Trends is a web analytical tool offered by Google. It allows analysis of various trends
4|Page
through the use of a range of dimensions and metrics such as; demographics, location,
acquisition, behaviour, time lapse etc. This tool will give us an insight into valuable
information surrounding climate change.
Netlytic
Netlytic is a text analysis tool which analyses text and social networks such as; Twitter,
Instagram and Facebook. It filters the comments, mentions and certain words from all posts
under the specific search of your choice to find useful information displayed in a range of
diagrams. We will be using two popular Twitter users, and the keywords “climate change” on
Twitter to understand people’s views on climate change.
Gephi
Gephi is a software platform which allows users to take data and create network
visualisations for analysis. Users can create the graph flexibly by customising layout and
colours, identifying communities clearly and changing the node size by degree etc. We
included data sets from keywords on Twitter which are “climate change” and, Trump and
Thunberg’s Twitter mentions. Therefore, each node is a representation of a Twitter user with
a tweet towards Trump or Thunberg. The network visualisation will assist us in
understanding the interactions and differing opinions.
Discussion
Google Trends: “Climate Change:
Interest Over Time:
These results have been filtered to see how frequently climate change has been searched
throughout the year. Since October 2018 to September 2019, the interest has dramatically
increased due to the world-wide awareness we are currently facing. During this timeframe, the
interest for this topic was steady and remained stationary for a long period of time until
September 2019. Therefore, the public are more aware of this issue and are eager to discover
more information regarding it. In September 2019, the high peak was triggered by the global
climate strikes that took place from September 20th-27th. Millions of people attended the
strikes which forced people with all kinds of opinions of the topic to notice.
5|Page
Interest By Region:
The analysed results shown identify the top regions of interest in regard to ‘climate change’.
It’s observed that Fiji is first in the list followed by Ethiopia and New Zealand coming in fourth.
Within the past 12 months, Fiji has projected interest in over 100 searches with climate change,
Ethiopia has approximately 25 and New Zealand has expressed 18. This provides a high
indication that Fiji has a greater concern and interest with this topic compared to other countries
displayed due to its high search usage. This may be due to them having more risk or concerns
with climate change therefore, the necessity to investigate more on the topic.
Related Topics and Related Queries:
Related topics and queries displayed shows users searched for the term ‘climate change’ also
searched the topics presented in the table. For example, the top of the list shows the most
popular topics. The ‘rising scale’ shows these related topics have a big increase in the
frequency of a search. The results that are marked (breakout) mean these topics are new and
have had very few searches. Compared to the bottom of the list, which show how many
searches these related topics have. It is interesting, that Thunberg is mentioned three times out
of the top five indicating that she is a popular activist towards climate change and possesses an
impact.
Text Analysis:
In order to conduct a text analysis around peoples differing thoughts and opinions on the
climate change crisis, we decided to compare two people with extremely different thoughts on
the topic; Trump and Thunburg.
6|Page
Word cloud:
Trump
(Figure 1)
Figure 1 displays the top 50 words that are tweeted in association with Trump. The words are
different sizes to show the relative frequency compared to others, and a number next to the
word indicates how often the word has been used. After analysing, it’s evident the top 3 most
frequently used words are, @realdonaldtrump, rt, and @cnn. It also clearly shows that in the
top 50 words there are no words present in relation to climate change or the environment. As
climate change is a crucial and controversial issue we were expecting one of the most powerful
people in the world to address concerns/solutions via Twitter.
Thunburg
(Figure 2)
Figure 2 shows the top 50 words that have been tweeted in association with Thunberg. It’s
evident that words such as ‘climate’, ‘humanity’ and ‘ecological’ have been used greatly as
they are bigger in font size. It is interesting, that a 16-year-old girl has more interest in the
environment and climate change, compared to Trump who holds power and a large following.
It is observed, that Thunberg has been connecting with many other Twitter users thus, she is
willing to communicate with others around the issue and spread awareness.
7|Page
Sentiment Analysis:
Trump:
(Figure 3)
(Figure 4) (Figure 5)
Figure 3 shows Trump’s sentiment analysis, identifying that 48 of the tweets, retweets and
other interactions are about ‘feeling good’ whereas 31 are about ‘feeling bad’. Upon analysing,
good feelings and bad feelings, it’s identified that there is not much of a difference between
the number of words used. The most commonly used words that display feeling good are
‘good’, ‘great’, ‘funny’ and ‘wonderful’, which is displayed in figure 4. Figure 5 displays
words like ‘bad’, ‘tired’, ‘evil’, and ‘nasty’ in association with bad feelings. It is interesting
that Trump has an outstanding amount of ‘bad feelings’ associated with him, which is linked
8|Page
to the controversy around him being president and not caring about extreme issues such as
climate change.
Thunburg:
(Figure 6)
(Figure 7) (Figure 8)
Figure 6 of Thunberg’s sentiment analysis shows that 23 posts are ‘good feelings’ whereas
only 4 are related to ‘bad feelings’, much less than that of Trump. It’s shown that the words in
relation to feeling good are; ‘good’, ‘proud’, ‘successful’, and ‘great’, compared to those of
feeling bad, ‘tired’, ‘sad’, ‘awful’, and ‘dangerous. The words that indicate feeling bad provide
an insight into the way Thunberg feels about the way humans treat the environment. It is
important to notice that the good feelings are still dramatically larger than the amount of bad
feelings, indicating that besides her frustration with the population, she still intends to spread
awareness through positive language.
9|Page
“Climate Change” Keywords:
(Figure 9)
(Figure 10) (Figure 11)
These figures are showcasing the Twitter interactions involving the keywords “climate
change”. Figure 9 outlines that there were 73 posts related to feeling good whereas 11 posts
are linked to bad feelings. This indicates that the majority of posts created on Twitter towards
“climate change” is positive which could be due to people mostly mentioning the positive
actions undertaken towards the environment. Figure 10 visualises the particular words used
within the feeling bad posts. The two most common words are “terrible” and “dangerous”
followed by, “selfish”, and “evil”. Therefore, we can identify how people describe the
negative impact of climate change. Figure 11 portrays the words used that were linked to
feeling good. “Proud” was most popularly used followed by, “good”, “great”, and, “healthy”.
Thus, showcasing how people see the positive acts towards perhaps reversing climate
change.
10 | P a g e
Gephi Network Visualisations:
Thunberg:
No. of nodes: 1180 No. of edges: 2706
The diagram displays a network visualisation of Thunberg’s data set. The network is filtered,
highlighting the important nodes, the intensity of the nodes and links through different colours,
followed by a range of clusters. “Community clusters form around popular topics. These
networks develop when few dominant hubs (users) with its audience, influence, and sources of
information” (Khan, 2018). In the visualisation, Thunberg is a climate change advocate
therefore, she is spreading global information that is attracting coverage from different outlets.
This forms a variety of different sized groups called “clusters” who share similar connections
to Thunberg. The number of edges is large however, indicating there are many nodes with 0
connections.
11 | P a g e
This diagram is filtered by degree, ranging from 20-546. Each of these nodes have more than
20 connections within the network, making them of higher importance within the network.
There are some notable users in this visualisation who are celebrities and verified users with
millions of followers. The status of these users indicate their importance not only to Thunberg,
but to the public’s opinion on climate change. There are a few clusters from different
modularity classes which suggests there are various highly-interconnected groups. The strength
of these, along with the influence ability of the users make the impact of their opinions on
climate change relatively strong. Thunberg is known for her climate change activism which
explains the various important connections in her network.
12 | P a g e
Trump:
No. of nodes: 1345 No. of edges: 1895
This network visualisation represents Trump and the Twitter users associated with him. The
network indicates that Trump connects to a large scale of people via Twitter, meaning that if
he was to positively activate for climate change it would reach a lots of people. Although, he
chooses to show negativity towards those who advocate for environmental change. There are
some obvious clusters within the network which indicates a communication path between
nodes. The components of the cluster are connected through the same or similar tweets. We
assume that these nodes have clustered due to them sharing similar feelings toward Trump.
13 | P a g e
This graph is filtered by degree, ranging from 20-788. Each of these nodes have more than 20
connections within the network, making them of higher importance within the network. The
importance of these nodes indicates that they are actively involved with tweeting to and about
Trump. These connections lack importance within Trump’s network compared to Thunberg’s
which indicates that people do not tend to communicate with or about Trump in terms of
climate change as it’s not important to him.
Thunberg, Trump and “climate change” keywords:
14 | P a g e
No. of nodes: 1345 No. of edges: 5101
This visualisation displays a combination of Twitter data from Trump, Thunberg and the
keywords “climate change”. There is high density in the central areas with five distinctive
communities (modularity class) surrounding. These communities show high interconnection
with one another which means they share similar opinions. The purple community which
Trump plays a large role in is more spread out and less clustered, indicating his subnetwork is
weakly connected. The green community which Thunberg plays a large role in, is tightly
clustered, which indicates those in this modularity are strongly connected to each other. This
shows how Thunberg and her subnetwork are strong and impactful within their community.
The biggest nodes are highlighted with their size and label size and represent Trump and
Thunberg, with Trump’s node being larger, therefore highly connected within the network.
Thus, Trump has a larger influence on public perceptions of climate change. His Twitter
holds the potential to make a big impact on the opinions of other users, which could have
negative effects on the Earth as he advocates against climate change. Hence, more people
disregard climate change as an issue and our planet will continue to deteriorate. Thunberg
still holds a great amount of influence within her network and as she continues to be an
activist, her network will grow.
Conclusion
Our investigation discussed the roles of two prominent figures in the climate change debate,
Trump and Thunberg and how their opinions differ on the issue. During the investigation,
Google Trends established the amount of use the term ‘climate change’ had over the past year.
It found that interest in climate change has risen significantly over the past year, with
Thunberg’s speech at the United Nations and global climate strikes causing a spike in
September. Netlytic allowed us to see how Trump and Thunberg were interacting with the
online community on the topic of climate change. This provided the visualization of the
common words and phrases used by either party and the meaning behind them. The findings
showed that the common words used by Thunberg fell into the ‘feel good’ category than those
used by Trump. Gephi helped visualize the connection between the influencers and their
respective communities. It found that Trump’s influence on the climate change topic was
stronger than Thunberg’s, therefore, more people could be disregarding climate change as an
issue.
Thus, Trump is still in a higher position in terms of influence than Thunberg, but this may
change as the support for the idea that climate change is a problem further grows and activists
like Thunberg gain more attraction in mainstream media.
Word Count: 3,000
15 | P a g e
References
An, X., Ganguly, A. R., Fang, Y., Scyphers, S. B., Hunter, A. M., & Dy, J. G. (2014).
Tracking climate change opinions from twitter data. Workshop on Data Science for Social
Good.
http://cobweb.cs.uga.edu/~squinn/mmd_s15/papers/KDD_Twitter_ClimateChange.pdf
Bergquist, P., & Warshaw, C. (2019). Does Global Warming Increase Public Concern
about Climate Change? The Journal of Politics, 81(2), 686-691.
https://www-journals-uchicago-edu.ezproxy.waikato.ac.nz/doi/pdfplus/10.1086/701766
McCright, A., & Dunlap, R. (2011). Cool dudes: The denial of climate change among
conservative white males in the United States. Global Environmental Change, 21, 1163-
1172. https://doi.org/10.1016/j.gloenvcha.2011.06.003
Khan, G.F. (2018). Creating value with social media analytics: Managing, aligning,
and mining social media text, networks, actions, location, apps, hyperlinks, multimedia, &
search engines data. Washington: CreateSpace.
Lee, T., Markowitz, E., Howe, P., Ko, C., & Leiserowitz, A. (2015). Predictors of
public climate change awareness and risk perception around the world. Nature Climate
Change, 5, 1014-1020. https://doi.org/10.1038/NCLIMATE2728
16 | P a g e
