Next Article in Journal
Greenspace and Atopic Sensitization in Children and Adolescents—A Systematic Review
Previous Article in Journal
Assessment of Health-Related Quality of Life between People with Parkinson’s Disease and Non-Parkinson’s: Using Data Drawn from the ‘100 for Parkinson’s’ Smartphone-Based Prospective Study
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Understanding #WorldEnvironmentDay User Opinions in Twitter: A Topic-Based Sentiment Analysis Approach

1
Department of Business Economics, Faculty of Social Sciences and Law, Rey Juan Carlos University, Paseo Artilleros s/n, 28032 Madrid, Spain
2
Institute for Global Law and Policy, Harvard Law School, Harvard University, Cambridge, MA 02138, USA
*
Author to whom correspondence should be addressed.
Int. J. Environ. Res. Public Health 2018, 15(11), 2537; https://0-doi-org.brum.beds.ac.uk/10.3390/ijerph15112537
Submission received: 22 September 2018 / Revised: 4 November 2018 / Accepted: 9 November 2018 / Published: 13 November 2018
(This article belongs to the Section Digital Health)

Abstract

:
The main objective of this exploratory study is to identify the social, economic, environmental and cultural factors related to the sustainable care of both environment and public health that most concern Twitter users. With 336 million active users as of 2018, Twitter is a social network that is increasingly used in research to get information and to understand public opinion as exemplified by Twitter users. In order to identify the factors related to the sustainable care of environment and public health, we have downloaded n = 5873 tweets that used the hashtag #WorldEnvironmentDay on the respective day. As the next step, sentiment analysis with an algorithm developed in Python and trained with data mining was applied to the sample of tweets to group them according to the expressed feelings. Thereafter, a textual analysis was used to group the tweets according to the Sustainable Development Goals (SDGs), identifying the key factors about environment and public health that most concern Twitter users. To this end, we used the qualitative analysis software NVivo Pro 12. The results of the analysis enabled us to establish the key factors that most concern users about the environment and public health such as climate change, global warming, extreme weather, water pollution, deforestation, climate risks, acid rain or massive industrialization. The conclusions of the present study can be useful to companies and institutions that have initiatives related to the environment and they also facilitate decision-making regarding the environment in non-profit organizations. Our findings will also serve the United Nations that will thoroughly review the 17 SDGs at the High-level Political Forum in 2019.

1. Introduction

New technologies allow users to express their opinions in the so-called social media, i.e., online platforms where users create communities around a topic of interest [1]. Owing to these new technologies, it is now possible to obtain information, carry out research on the expressed sentiments, or extract the key factors, the specific factors that receive special interest from a specific community, that serve both the users and the institutions that develop the agenda of actions vis-à-vis these problems [2,3].
One of such topics is the environment, this particular concern for the environment and how it can affect people’s health is well grounded [4,5]. Ever since pioneers of sustainable development began to study the impact of human activities on the environment [6,7] several decades ago, the global situation has progressively worsened. By now, it has reached the point where natural resources are being consumed at the rate which is 40% higher than what the Earth can endure according to [8]. The increase in this kind of events, is linked to climate change and the environment in general, so that research like this, can help eradicate this type of events by providing information and analysis methodologies for future research or studies in this field.
Furthermore, nowadays, it is essential to not only manage the resources but also to ensure that these resources will be available for future generations and to minimize the impact of human activities on the environment and public health [7,8,9].
Taking into consideration that sustainable management of resources is of paramount importance [10], institutions and their leaders are creating principles named Sustainable Development Goals (SDGs), to ensure the viability of the planet.

1.1. Sustainable Development Goals

In a meeting convened by the United Nations Organization (UN) that took place in 1992, world leaders established the principles of the Rio Declaration and the Agenda 21, a document of principles which laid the foundations for the sustainable development of the 21st century. The meeting was attended by more than 150 heads of State. It was in 2000 when the United Nations Educational, Scientific and Cultural Organization (UNESCO) created the Millenium Development Goals (MDGs).
In this new approach, the agendas of the United Nations Development Programs (UNDP) have been collected, along with those of the United Nations Environment Program (UNEP), World Health Organization (WHO), United Nations Children’s Fund (UNICEF), UNESCO, as well as other development agencies [11].
Since then, the interest in the environment and public health have become the main concerns for society [6]. From that moment, people started to be considered as a fundamental part of sustainable development and, consequently, governments and institutions started to invest in them, in their health, and in the environment that supports their existence [7], as summarized in the Agenda 21 promulgation: “Human beings are the center of concerns for sustainable development; they have the right to a healthy and productive life in harmony with nature” [12].
After the meeting, numerous international institutions have sought to give continuity to the principles that bring together the environment, health, and sustainable development, such as the one organized by the World Health Organization [13].
All these developments at the national and international political levels have promoted a holistic perspective on health, economy, and environment. Along with economy [14] and decision-making [15], health is now becoming a global concern for almost all sectors of the society.
With the global aim of promoting sustainability, public health, and the environment, in 2015, the United Nations presented the SDGs, where 17 principles of sustainable development based on Agenda 21 were formulated emphasizing equality, human rights, and non-discrimination of people [3,16,17,18]. Some of the principles include Climate Action, Life on Land, or Clean Water and Sanitation, were formulated [19,20,21] (see Table 1 for the complete list). Yet, to what extent does the society share the concerns about environment and public health of the SDG leaders?
Simultaneously with the development and implementation of the principles of sustainable development, the use of new Information Communication Technologies, such as social media and mobile-based communications, is revolutionizing the way users and businesses communicate with each other [22,23]. Here, the use of inadequate communication strategies is responsible for the low impact that efforts related to the environment have had [24]. In this sense, the adequate use of technology can facilitate the communication tasks about the environment, and, by doing so, help to achieve sustainable [1] and effective management [25]. In addition, the use of new information and communication technologies makes it possible to get a comprehensive picture of user concerns and to identify the most important factors involved therein [26,27].
Consequently, the main objective of this exploratory study is to identify the social, economic, environmental and cultural factors related to the sustainable care of both environment and public health that most concern Twitter users. The results of the analysis enabled us to establish the key factors that most concern users about the environment and public health.

1.2. Twitter as a Social Communication Platform to Measure Public Opinion

In recent years, social media has emerged as a virtual space where users can express their concerns about both the environment and health and other issues of public interest [28,29,30,31]. Actually, the use of social media to develop daily activities is very high worldwide, being an essential source of information on the communications and opinions of users on the Internet [24]. Every minute, some 473,000 tweets are published, 3,788,140 searches are performed on Google, and 18,055,555 text messages are sent [24].
Among social media, the one that most favors users’ sharing their opinions is Twitter [25]. With its 336 million active users, Twitter is increasingly useful for the extraction of factors from public opinion in social media [28,32]. This social network serves as an optimal starting point, since, unlike Facebook that offers limitations with private or semi-private profiles, Twitter facilitates open communication between users and allows them to share their opinions in a maximally accessible way publicly, main advantage with respect to other social networks that are private [33].
Twitter allows users to create and share opinions on the spur of the moment. It also allows users to interact with each other either by following other users, or by visiting their profiles and sharing the same hashtag about a common topic of interest. A hashtag is a tag that is represented by a pad (#) and that allows users to group all the published tweets on a given topic [34]. The comments with a hashtag can be visualized together; their retweets can be visualized as well. A retweet is a direct sharing of a comment written by other users, as what they write is aligned with the opinion of the person who does the retweet so that the comment appears directly in his/her profile. The most used hashtags become topics of general interest, or trending topics [35].
Table 2 was created by the authors based on Thapa [36] and Pena [37] to show the most used hashtags when earthquakes were produced in different parts of the world. In the case of the Mexican earthquake, the hashtags #sismo and #FuerzaMexico were the most frequently used, highlighting thus the trending topics in that country throughout the year 2017. Regarding the earthquake that hit Haiti in 2010 and shocked the world, #Haiti was the most shared hashtag throughout the world during that year. Furthermore, with respect to the earthquake in Nepal, during the month following the catastrophe, 33,610 tweets were published with the hashtag #nepalearthquake [36]. These figures show how the comments are grouped with respect to a topic around a series of hashtags shared by other users and can be studied to obtain information [30].
Among the first to study Twitter as a new means of communication were Java et al. [31] who highlighted that this social network is a space to share and search for information, as well as a space where users with similar interests can meet and form communities. Thereafter, Twitter has become increasingly used for research purposes [25], including the studies on public opinion [30]. In addition, Twitter can be used by companies to identify the key factors about environment and public health that can help them improve their management and provide user insights as it is explained in the discussion section of this study [32,33]. Unlike other social media such as Facebook, the growth statistics of Twitter are positive, the number of its active users is growing every month based on Aggarwal [34]. Therefore, Twitter is a very relevant social network [38].

2. Literature Review

Since the formulation of the SDGs as reference points for sustainable development, several studies took them as a reference [9,39]. This is due to the need of quantitative measures that the governments can rely on to assess their performance with respect to the challenges posed that range from natural resources management to pollution control.
When researchers aim to get information for governmental decision-making about a certain topic, one the means they use is to assess the opinion of individuals or users in social media [3,32].
Even there is a growing community in social media like Facebook or Twitter and each time more individuals have a voice on these networks about such topics as the environment, this media has shown its weaknesses as a single means for policy-making as not all the individuals are represented by social media users [40].
However, there are numerous studies have used Twitter and hashtags to identify topic-based feelings of the population regarding a specific issue related to the environment and public health [32,38]; other studies, e.g., those investigating public opinion in social media after a natural disaster, analyzed the keywords extracted from the comments [41]. Table 3 gives an overview of the object of our research based on previous important related work.
Among other relevant studies is the research by Woo et al. [41] that examined how Twitter users used hashtags to communicate during an earthquake or a flood since they could not make phone calls. Also, Chisholm and O’Sullivan [44] analyzed the hashtag #characterbuildings used to express Twitter users’ views on sustainable housing in New Zealand. Likewise, Saura et al. [2] used published tweets to identify social, economic, environmental and cultural factors that allow hotel establishments to improve decision-making and management of sustainability.
Tlebere et al. [44] developed a model of social awareness about the environment based on social media with the aim of providing people with the knowledge, skills, and attitudes necessary to reduce adverse environmental impact. Furthermore, Ekenga et al. [28] used the social platform to measure public sentiment about environmental health, obtain information on the issues that most concern the community, and communicate with them about environmental risks.
In the present study, following Chisholm and O’Sullivan [32] and Palomino et al. [38], we will focus on public opinion about the environment through the analysis of the hashtag #WorldEnvironmentDay in Twitter. #WorldEnvironmentDay was the most used hashtag to express opinions about the Environment on the day of the environment so that it joins the opinions of Twitter users about the topic. Methodologically, we will follow Saura et al. [2] by using sentiment and textual analyses. Firstly, we have downloaded tweets that used the hashtag #WorldEnvironmentDay on the respective day. As the next step, sentiment analysis was applied to the sample of tweets to group them according to the expressed feelings. Thereafter, a textual analysis was used to group the tweets according to the SDGs, identifying the key factors about environment and public health that most concern Twitter users. To this end, we used the qualitative analysis software NVivo Pro 12 (QSR International, Melbourne, VIC, Australia). The results will help identify the key factors about environment and public health that most concern the public (Twitter users). Thus, the main objective of this exploratory study is to identify the social, economic, environmental and cultural factors related to the sustainable care of both environment and public health that most concern Twitter users. Therefore, we have structured this study in the following parts: Introduction, Materials and Methods, Methodology, Results, Discussion and Conclusions.

2.1. Sentiment Analysis Approaches for Social Network Analysis

Sentiment analysis is an increasingly used methodology to identify the emotional burden contained in the messages published on a given topic [28,45]. Deriving from the use of Big Data, it is used to conduct research on the impact of events on social media, to measure the opinion of products and services, as well as to understand communication in online environments [38,46]. In essence, sentiment analysis involves downloading topic-based comments published on Twitter. After that, we can analyze user opinions of the users expressed in these comments using the models developed with machine learning [47].
Sentiment analysis can be combined with other technologies to extract the most important factors under study [48]. For instance, Pak and Paroubek [49] developed several different methodologies that can be used in Twitter analysis. In general, the analysis takes into account the connotations of the frequently recurring words, which signal different feelings and can be classified into positive, negative, or neutral. For example, Palomino et al. [38] used sentiment analysis to identify public opinion and study the related discourse to understand the impact of the natural environment on people’s health and well-being. On the other hand, Ekenga et al. [28] conducted a topic-based sentiment analysis of Twitter comments and demonstrated that this social media platform serves to measure public sentiment about environmental health, to obtain information on the issues that most disturb the community, and to communicate about environmental risks. All in all, previous research has demonstrated that social media are a good platform to perform this type of analysis [28,49]. Although some of the challenges of sentiment analysis are those related to the context in which content is analyzed, sarcasm or ironies, this technology can be trained to mitigate these effects and diminish these weaknesses. Precisely, with the use of machine learning and data mining processes we can anticipate ironies and sarcasms to anticipate this problem. In this research, those tweets that have been classified as sarcasm, ironies or in unknown contexts have been classified as neutral. This fact, is still a challenge and a limitation to improve the development of new technologies in research and not to validate this type of process that allows much more advanced analysis and with higher volumes of data than can be analyzed by the human being [3].

2.2. Textual Analysis

Textual analysis is a type of qualitative analysis that is applied to a text with the aim of grouping the concepts under study into “nodes”. This approach can be applied to an event, a company, or any other object of study [50]. The software most used to perform this analysis is NVivo (QSR International, Melbourne, Australia) [2,51]. The outcome of the method is grouping of the concepts in nodes that are very useful to perform exploratory analysis [52]. This approach has a greater descriptive capacity than if the software had not been used [53,54].
The nodes are distributed in different hierarchical levels. At the first level, there are concepts that are conceptually independent of each other. At the second level, there are the branches that start with each of the nodes and follow a hierarchical organization. Finally, there are the indicators that can be extracted from the research results and that are related to the aim of the specific study [52]. One of the relevant studies where textual analysis was used to identify factors related to the environment was conducted by Saura et al. [2]. Each of the three nodes identified by the authors corresponds to negative factors, positive factors, and neutral factors.

3. Conceptual Framework and Hypothesis Development

As already highlighted in the previous sections, social media have been increasingly used to obtain information about public opinion, sentiment, and key factors related to various objects of study. For instance, Chisholm and O’Sullivan [32] used hashtags to identify Twitter user sentiment. Similarly, Palomino et al. [38] used the social network to identify public opinion through the analysis of the tweets of a hashtag and to study the feelings expressed in the discourse related to nature-deficit disorder and other nature-heath concepts. In this sense, to validate that it is possible to identify sentiments on Twitter related to a specific topic and based on the previous research, the following hypothesis can be formulated:
Hypothesis 1 (H1).
The hashtag #WorldEnvironmentDay serves to identify the feelings of users about the environment.
Furthermore, Woo et al. [41] performed the analysis of public opinion through the study of the tweets published in relation to a natural disaster in Korea. Based on the relevant comments, the authors identified how Twitter users’ sentiment changed and how topic-based keywords related to the natural disaster appeared. Likewise, in their analysis of comments on Twitter profiles of hotels, Saura et al. [2] identified the most important factors related to the environment. Consequently, with the objective of identifying important factors for the environment, we propose the following hypothesis based on the similar above-mentioned research:
Hypothesis 2 (H2).
The analysis of the communications of users who use the hashtag #WorldEnvironmentDay can be used to establish the key factors related to the environment.
Next, the starting point for the studies by Messerlin et al. [26] and Browman and Stergiou [9] was, among other topics of public interest, the SDGs with their 17 principles related to the environment to identify the key factors about environment and public health. On the other hand, Kumar et al. [3] selected the SDG related to health and the environment for their analysis. Therefore, with the aim of being able to link the important factors related to the SDG established by the United Nations, we propose the following hypothesis:
Hypothesis 3 (H3).
The communications of users on Twitter through the hashtag #WordEnvironmentDay during the respective day are related to Sustainable Development Goals (SDGs).

4. Methodology

The first of the two methodologies employed in the present study is sentiment analysis of the tweets published by users on Twitter with the hashtag #WorldEnvironmentDay. Following Palomino et al. [38] and Saura et al. [2], firstly, we developed a sentiment analysis. The second methodology was textual analysis, in which the qualitative analysis software NVivo Pro12 was used to categorize the tweets into positive, neutral, and negative categories in order to identify key factors related to each of these three categories. The results make it possible to identify the factors related to the sustainable care of the environment that are most relevant to public opinion in Social Media and organize them with regard to SDGs and the key factors about environment and public health. To this end, we relied on the methodological process proposed by Saura et al. [2] and Ekenga et al. [28].

4.1. Sample

In two previous studies, Bologna and Hayashi [53] and Palomino et al. [38], sentiment analysis was applied to two Twitter samples, one with 2000 tweets and another with 10,000 tweets, respectively [54,55,56]. On the other hand, Palomino et al. [38] retrieved 6333 tweets related to the hashtag #getoutside. Following these previous studies, the sample in the present study was composed of 5873 tweets that used the hashtag #WorldEnvironmentDay and were published in English on 5 June 2018 [57]. At first, the sample included 9467 tweets, but, after the process of cleaning up the database, the final sample was reduced to 5873 tweets. Retweets using #WorldEnvironmentDay were removed because it is considered duplicate content that does not enrich the database [38]. Of the users participating in the study, 47.5% were male, 52.5% were female, and no other gender was detected (as provided by the Twitter API) [58,59,60]. In addition, following Saura et al. [2], the sample of tweets was validated according to the following criteria [61]:
  • Active Twitter profile (profiles without activity for three months prior to the use of #WorldEnvironmentDay were deleted)
  • Twitter user profile with a profile photo and a cover
  • Retweets from the same tweet about #WorldEnvironmentDay were removed (i.e., considered as duplicate content)
  • Only public profiles and tweets in English using #WorldEnvironmentDay on 5 June 2018, were included
  • Tweets should have been at least 80 characters long with spaces and use the tag #WorldEnvironmentDay; therefore, tweets without the “#” or a wrong label like #WorldEnvironmentDay2018 were omitted.
The tweets came from all over the world. Table 4 specifies the weight of each country from which relevant tweets were made. The three most active countries in terms of number of tweets were the United States, United Kingdom, and India. The countries with the smallest number of tweets were Malaysia, Mexico, and Belgium.

4.2. Data Collection and Extraction

For the data collection and extraction process, we connected to the public Twitter API on 7 June 2018, two days after the #WorldEnvironmenDay for which we use Python software 3.7.0 (Python, Wilmington, DE, USA) in the version for MAC. This was made to ensure that all the countries of the world would have had the opportunity to publish tweets with respect to #WordEnvironmentDay for the time difference and within the capabilities of the Twitter API that only allows data to be extracted during the last 7 days of activity. The day of #WordEnvironmentDay is the day chosen by the United Nations and other global associations to highlight the care of the planet through the environment. Other purposes include carrying out social actions that demand the care of issues directly related to #WordEnvironmentDay such as global warming or pollution, and encourage new initiatives that improve the care of the planet [6].
As specified in Section 4.1, after cleaning up the database, the sample was composed of 5873 tweets. To this database, we applied the algorithm developed in Python by MonkeyLearn (San Francisco, CA, USA) [62] after connecting to its API. This was done after training the machine with data-mining processes and subdividing it into positive, negative and neutral feelings. In the datamining development process, a total of 732 samples were trained. The training of these samples of tweets was done through the MonkeyLearn application and its interface linked to the Sentiment Analysis algorithm until reaching the probability percentage of >0.650. This training of the algorithm was carried out by the authors of the research under criteria linked to the identification of ironies, sarcasms and contents related exclusively to the objective of this research. In the entire process, those contents that were not related to #WordEnvironmentDay were discarded from the sample and training. Subsequently, the different divisions of the databases according to sentiment were processed using NVivo Pro 12 Software [63] at different stages in which the tweets were categorized according to the following three nodes according to the feelings expressed by the tweets: Positive (N1), Neutral (N2), and Negative (N3).

4.3. Sentiment Analysis and Textual Analysis

With regard to sentiment analysis, we first connected to the algorithm and trained it with data-mining processes using a sample of tweets related to the environment and sustainability until we managed to increase the significance and possibility of algorithm prediction a >0.650 (probability percentage), which is an indicator that measures the average success of a machine when using machine learning techniques [2]. The process of possibility of prediction of the algorithms that work with machine learning is an indicator that measures the effectiveness of analysis of the machine according to the feeling and can be identified in terms of the average of the database or the analysis of a single tweet [51].
The probability percentage is a measure that sets out the accuracy and recall of samples in each category. This percentage is the result of the classification success obtained by the Support Vector Machine (SVM) algorithm that works with machine learning and that was trained with data mining to perform the sentiment analysis. This percentage defines the total average success of the algorithm in the review’s classification. Similarly, it should be mentioned that we used an algorithm based on SVM typing. Supervised learning is the most popular category of machine learning algorithms. The disadvantage of this approach is the fact that, for each training example, we must provide the correct result until the algorithm gets a correct percentage of success. SVM algorithms are a non-probabilistic model that uses a representation of text examples as points in a multidimensional space. These examples are mapped so that the examples of the different categories (feelings) belong to distinct regions of this space. Then, the new texts are mapped on this same space and predicted as belonging to a category according to the region in which they are located [55,64,65].
Next, the different databases according to feelings were processed using NVivo Pro 12 at different stages in which the tweets were categorized into the following three nodes: Positive (N1), Neutral (N2), and Negative (N3). This process of entering the data in Nvivo is manual, although the databases are already divided in their feelings. Subsequently, the researchers proceed to make the structure of the nodes and filter the database eliminating those words as connectors, prepositions or articles and their plural forms. Then, the nodes are predefined data containers grouped according to their characteristics. It should be noted that the design and development of nodes is a norm to analyze pure data and to achieve the highest descriptive and exploratory quality possible [2]. In this sense, a relevant indicator in the analysis with NVivo Pro 12 is the weighted percentage, which represents the weight of the indicators grouped into nodes according to the times they are repeated, thus being the weight of the nods in terms of the total data that exist in the database. To calculate the weighted percentage, we used NVivo Pro 12 with the following formula (see Equation (1)):
K = ∑ki/ni = {1, …, n} n = [1, 25]
In the formula, K is found using a query that allows the program to search the text. The behavior of each of the words and for each tweet can be seen. Therefore, the K value was found for the Hashtag #WorldEnvironmentDay. In this way, the average K for all the tweets was calculated in order to obtain the global value [2].

5. Results

As mentioned above and as can also be seen in Figure 1, as the first step of the methodological process, we performed data extraction and data collection from the Twitter API. After completing these processes and debugging the data, we obtained the database of n = 5873 tweets.
After applying sentiment analysis, all analyzed tweets that used #WorldEnvironmentDay were divided into Positive (P, n = 1243), Neutral (X, n = 2687), and Negative (N, n = 1943). The total mean probability coefficient of the sentiment analysis was 0.719 for P, 0.651 for X, and 0.802 for N (see Table 5). As can be seen in Table 4, the average total probability coefficient was 0.724, and it was greater than >0.650. This value needs to be considered in the analysis since it indicates the probability of success when classifying the tweets by the tool used.
After obtaining the three databases according to the expressed feelings, in order to identify the factors related to the environment, the data related to each feeling were added into different nodes using NVivo Pro 12. Therefore, the structure had three different nodes, N1 (positive), N2 (neutral), and N3 (negative). On these independent nodes, textual analysis was carried out. Textual analysis subdivided the pure data into content categories related to the environment according to the feelings expressed by the tweets. Therefore, the division of the databases into nodes yielded the following distribution of tweets in the three nodes: N1 = 1243 tweets, N2 = 2687 tweets, and N3 = 1943 tweets (see Figure 2).
For the structure of the nodes N1, N2, and N3, we also took into account the SDGs related to the environment and public health. These served as categorizers according to the feelings shown by the users in the tweets related to #WorldEnvironmentDay (for an example, see Table 6).
Likewise, the hashtags related to #WorldEnvironmentDay that refer to different actions and programs that support sustainable development, the environment, and public health were identified (see Table 7).
Specifically, one can identify topics related to climate change, the elimination of materials such as plastic and initiatives against it, programs related to oceans and fish farms, as well as other initiatives to promote the environment and sustainability in general.
The use of textual analysis yielded the following grouping of tweets into nodes according to the expressed feelings [64,65,66,67]. Results have been obtained grouped in Nodes according to feeling.
Table 8 shows the percentages corresponding to the weighted percentage indicator, as well as the roaming of the words representing the nodes and the concepts related to each of the indicators identified by textual analysis. Likewise, each of the nodes represents the sentiment of the topics related to the SDGs identified in the database of tweets corresponding to #WorldEnvironmentDay. Therefore, we can observe the category of the identified indicators, the roaming number, similar factors that give shape to the node, as well as its weighted percentage in the entire database. As a result of Sentiment Analysis and Textual Analysis processes directly from the research data, factors indicated in Table 8 are identified.

6. Discussion

As demonstrated in previous research, public opinion in social media and associated feelings can be measured with respect to a specific research topic in the social network Twitter. In the present study, we have demonstrated that we can identify the feelings contained in the indicators related to the key factors about environment and public health, and closely linked to the SDG, in Twitter.
As suggested by our results, there is a greater negative concern with respect to the environmental indicators included into the following SDGs: Climate Action (N4—Climate Change), Clean Water and Sanitation (N4—Water), Life on Land (N4—Forest), Responsible for Production (N4—Pollution and Biodiversity), and Sustainable Cities and Communities (Massive Industrialization). Therefore, it can be concluded that public opinion in social media is negative regarding the environment and factors such as climate change, problems with water in terms of pollution, deforestation or massive tree felling, the pollution that drives climatic changes and increases the risk of weakening of the atmosphere, biodiversity and its growing scarcity, and, finally, massive industrialization that eliminates populations and affects poor communities that face difficulties in industrialized cities.
The negative factors that were detected may be the most useful for non-profit organizations and groups that fight for sustainable development, since public opinion in social media, through its negative evaluation, demonstrates its concern with respect to a specific issue. In our case, this topic is environmental protection materialized in the SDG. In addition, if we link the topics of the SDG with the indicators related to the environment published around #WorldEnvironmentDay, we can relate each of them to the following SDGs: Good Health and Well-being, Clean water and Sanitation, Affordable and Clean Energy, Industry Innovation and Infrastructure, Sustainable Cities and Communities, Responsible for Production, Climate Action, Life Below Water and Life on Land.
Our results also suggest that, in terms of positive indicators related to the environment, public concerns are grouped around public health, as well as clean energy and sustainability. Furthermore, in terms of neutral indicators related to the environment, there are indicators related to sustainable cities, development of sustainable resources, and promotion of a healthy lifestyle. Therefore, the factors perceived as positive are those related to the right actions taken about the environment while neutral or negative ones need to be improved. In addition, the results of the present study demonstrate that public opinion on Twitter can be measured with respect to impact issues, thus allowing to identify the key factors associated with a given topic, which can ultimately be evaluated against the agendas proposed by policy-makers. In the present study, we have shown this by analyzing the tweets with the hashtag #WorldEnvironmentDay and linking them to the SDGs and the key factors about environment and public health proposed by the United Nations.
Similarly, we have shown that Twitter is a tool that can be meaningfully used to measure sentiment regarding an emerging topic through the analysis of tweets linked to a hashtag. Therefore, social media can be regarded to be relevant venues enabling the analysis of public opinion in Social Media, which can ultimately facilitate decision-making of companies, institutions, and global communities that support sustainable development [64,65,66].
Finally, our results have demonstrated that the use of new emerging technologies, such as techniques based on machine learning and sentiment analysis algorithms developed in Python, can be useful to improve the understanding of the community of a global society, as well as to develop analysis strategies for large amounts of data.

7. Conclusions

In this context, new technological environments [68] and, in particular, social media, have become a reflection of the evolution of new technologies. In particular, social media have become a new channel used by people, social movements, political parties, companies, non-profit associations, or communities to express their opinion or concerns about a determined subject [2,38,69].
In the present study, focusing on the analysis of tweets with the hashtag #WorldEnvironmentDay, we have been able to identify the main factors that concern the global population with respect to the sustainable development of the planet, public health, and the environment [70,71,72].
The importance of our results is determined by the relevance of the analysis of public opinion in social media about the environment in social media. Also, considering the initiatives such as the United Nations’ development of the SDGs, the use of this technique to analyze the users’ feelings is of relevance to the scientific community and can serve to substantiate the findings obtained using other methodologies [2,38,53,72]. This fact precisely acquires relevance insofar as the contributions of this research are based on models and techniques that analyze a social network in which millions of users participate. Therefore, the use of these techniques and the results of studies of these characteristics can help to contrast studies already carried out or to identify new problems related to climate change [2,53,72]. As demonstrated by our results, the analysis of the sentiment expressed in the communications of Twitter users can help identify social, economic, environmental and cultural factors regarding their positive, neutral or negative charge [73,74,75,76]. Regarding Hypothesis 1 (H1), the results of sentiment analysis confirm that users’ feelings about the environment can be identified. This finding suggests that the Twitter platform can be used to measure the feeling of users with respect to a specific topic. Furthermore, our results also support Hypothesis 2 (H2), since textual analysis identifies the feeling of users about the environment through the study of the hashtag #WordEnvironmentDay [77]. Finally, Hypothesis 3 is supported as well, since, as has been demonstrated with the development of the methodological process, the indicators identified after textual analysis and sentiment analysis could be linked to the SDGs formulated by the United Nations.

7.1. Implications for the Industry

Taken together, the results of the present study can be meaningfully used by NGOs, social associations, or companies that support sustainable development programs to improve their policies and initiatives related to the environment and sustainability. Our findings can also be used by the United Nations and other institutions that see to develop or improve, where appropriate, the new SDGs in 2019. During this year, they will be reviewed again and taking into account the potential of social networks to measure public opinion on specific issues, this research can be used to improve the development of new SDGs.
In addition, policymakers can also use the results of this research to focus their future efforts on activities and actions aimed at improving the environment and combating climate change. In addition, they can confirm that social networks such as Twitter can be a good communication channel to publicize their activities and social projects on world-renowned days such as #WorldEnvironmentDay and the topics related to this event. In addition, with data offered by this study methodology, they can even improve the segmentation of their target audience in social networks to carry out concrete actions in certain countries that are concerned about the environment and that actively demonstrate this concern on Twitter.

7.2. Implications for Academics

The events that involve the planet in problems related to the environment have been the object of thorough research over the past decade [34,67]. This research uses relatively new methodologies, and its main innovation lies in the use of new technologies such as machine learning and datamining, and its connection with social networks such as Twitter that are known worldwide and that bring together a multitude of users around specific themes [2].
Academics can use this research as a bibliographic source not only for the environment but also for the increase of references to methodologies such as sentiment analysis applied to social networks, public opinion on social networks and the development of machine learning. In addition, within the environmental sector, the present study refers to specific research related to environmental research based on emerging technologies. In this context, we conclude one of the benefits of applying a sentiment analysis process is that there are no biases in the classification by sentiment since this is done by an algorithm and not by a researcher manually. The present study developed a methodology based on innovation that can be used in other studies with larger samples in which it is necessary to automate processes. Academics can use this study, therefore, to find references within this relatively new category of research and emerging technologies.
The indicators established in the present study convey public opinion in social media on such important issues, as global climate change, deforestation, water pollution, and unsustainable building overcrowding [72] and can be used in future research as the object of study. In addition, researchers can use the results of this study in future research to improve the methodology and focus on sectors based on technology and innovation such as Twitter and the opinions of users with respect to a specific topic in social networks.
The limitations of the present study are related to the time horizon analyzed, the limitations of the social network Twitter itself, as well as the number of users who made comments using the hashtag under study. Likewise, the limitations of the present study are also related to the level of accuracy of the algorithm trained with machine learning processes that identify the sentiments of the sample as well as the consequently establishment of the factors identified in this study.

Author Contributions

A.R.-M., J.R.S. and C.A.-A. conceived and designed the review; J.R.S. performed the methodology; A.R.-M. wrote the introduction, related work and hypothesis development sections; J.R.S., A.R.-M. and C.A.-A. analyzed the results; A.R.-M., J.R.S. and C.A.-A. wrote the paper.

Funding

This research received no external funding.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Adi, A. Sustainability on Twitter: Loose Ties and Green-Washing CSR. In Corporate Responsibility and Digital Communities; Palgrave Macmillan: Cham, Switzerland, 2017; pp. 99–122. [Google Scholar]
  2. Saura, J.R.; Palos-Sanchez, P.R.; Rios Martin, M.A. Attitudes to environmental factors in the tourism sector expressed in online comments: An exploratory study. Int. J. Environ. Res. Public Health 2018, 15, 553. [Google Scholar] [CrossRef] [PubMed]
  3. Kumar, S.; Kumar, N.; Vivekadhish, S. Millennium development goals (MDGs) to sustainable development goals (SDGs): Addressing unfinished agenda and strengthening sustainable development and partnership. Indian J. Community Med. 2016, 41, 1. [Google Scholar] [CrossRef] [PubMed]
  4. Brundtland, G.H. Our Common Future: World Commission on Environment and Development; Oxford University Press: Oxford, UK, 1987. [Google Scholar]
  5. Palos-Sanchez, P.; Martin-Velicia, F.; Saura, J.R. Complexity in the Acceptance of Sustainable Search Engines on the Internet: An Analysis of Unobserved Heterogeneity with FIMIX-PLS. Complexity 2018. [Google Scholar] [CrossRef]
  6. United Nations. Agenda 21: Results of the World Conference on Environment and Development; United Nations: New York, NY, USA, 1992. [Google Scholar]
  7. Graham, J.L.; Andreasen, A.R. Marketing Social Change: Changing Behavior to Promote Health, Social Development, and the Environment. J. Mark. Res. 1997, 34, 294. [Google Scholar] [CrossRef]
  8. Santiago Fink, H. Human-Nature for Climate Action: Nature-Based Solutions for Urban Sustainability. Sustainability 2016, 8, 254. [Google Scholar] [CrossRef]
  9. Browman, H.; Stergiou, K. Perspectives on ecosystem-based approaches to the management of marine resources. Mar. Ecol. Prog. Ser. 2004, 274, 269–303. [Google Scholar] [CrossRef] [Green Version]
  10. Global Footprint Network. Earth Overshoot Day. 2010. Available online: https://footprintnetwork.org/ (accessed on 21 May 2018).
  11. Kumar, S. Health in international development Agenda: Present, past and future. Indian J. Community Med. 2013, 38, 129–131. [Google Scholar] [CrossRef] [PubMed]
  12. United Nations. Agenda 21: The United Nations Programme of Action from Rio; United Nations: New York, NY, USA, 1993. [Google Scholar]
  13. World Health Organization. Health and Environment in Sustainable Development: Five Years after the Earth Summit; World Health Organization: Geneva, Switzerland, 1997. [Google Scholar]
  14. Mihi-Ramirez, A.; García-Rodríguez, Y.; Cuenca-García, E. Innovation and International High Skilled Migration. Eng. Econ. 2016, 27. [Google Scholar] [CrossRef]
  15. Miller, K.M.; Hofstetter, R.; Krohmer, H.; Zhang, Z.J. Measuring Consumers’ Willingness to Pay. Which Method Fits Best? GfK Mark. Intell. Rev. 2012, 4. [Google Scholar] [CrossRef]
  16. Messerlin, P. Trade and Trade Policy Issues in the United Nations Millennium Development Goals and the Sustainable Development Goals. No. 638. 2017. Available online: https://www.adb.org/sites/default/files/publication/223481/adbi-wp638.pdf (accessed on 23 June 2018).
  17. Litsios, S. Sustainable development is healthy development. World Health Forum 1994, 15, 193–195. [Google Scholar] [PubMed]
  18. Warford, J.J. Environment, health and sustainable development: The role of economic instruments and policies. Bull. World Health Organ. 1995, 73, 387–395. [Google Scholar] [PubMed]
  19. Wysokińska, Z. Millenium Development Goals/UN and Sustainable Development Goals/UN as Instruments for Realising Sustainable Development Concept in the Global Economy. Comp. Econ. Res. 2017, 20, 101–118. [Google Scholar] [CrossRef] [Green Version]
  20. Pedercini, M.; Zuellich, G.; Dianati, K.; Arquitt, S. Toward achieving Sustainable Development Goals in Ivory Coast: Simulating pathways to sustainable development. Sustain. Dev. 2018. [Google Scholar] [CrossRef]
  21. Vladimirova, K.; Le Blanc, D. Exploring Links Between Education and Sustainable Development Goals Through the Lens of UN Flagship Reports. Sustain. Dev. 2016, 24, 254–271. [Google Scholar] [CrossRef]
  22. Fuster Morell, M.; Espelt, R. A Framework for Assessing Democratic Qualities in Collaborative Economy Platforms: Analysis of 10 Cases in Barcelona. Urban Sci. 2018, 2, 61. [Google Scholar] [CrossRef]
  23. Hubert, M.; Blut, M.; Brock, C.; Backhaus, C.; Eberhardt, T. Acceptance of Smartphone-Based Mobile Shopping: Mobile Benefits, Customer Characteristics, Perceived Risks, and the Impact of Application Context. Psychol. Mark. 2017, 34, 175–194. [Google Scholar] [CrossRef] [Green Version]
  24. Davis, J.J. The Effects of Message Framing on Response to Environmental Communications. J. Mass Commun. Q. 1995, 72, 285–299. [Google Scholar] [CrossRef]
  25. Leslie, J.; Mochon, D.; Emrich, O.; Schwartz, J. What’s the Value of a Like? Social Media Endorsements Don’t Work the Way You Might Think. Harv. Bus. Rev. 2017, 95, 108–115. [Google Scholar]
  26. Miller, K.; Skiera, B. Economic Damage of Cookie Lifetime Restrictions. SSRN Electron. J. 2017. [Google Scholar] [CrossRef]
  27. Gupta, S. For Mobile Devices, Think Apps, Not Ads.(cover story). Harv. Bus. Rev. 2013, 91, 70–75. [Google Scholar]
  28. Ekenga, C.C.; McElwain, C.-A.; Sprague, N. Examining Public Perceptions about Lead in School Drinking Water: A Mixed-Methods Analysis of Twitter Response to an Environmental Health Hazard. Int. J. Environ. Res. Public Health 2018, 15, 162. [Google Scholar] [CrossRef] [PubMed]
  29. Rathod, T.; Barot, M. Trend Analysis on Twitter for Predicting Public Opinion on Ongoing Events. Int. J. Comput. Appl. 2018, 180, 13–17. [Google Scholar] [CrossRef]
  30. Palos-Sanchez, P.R.; Saura, J.R.; Reyes-Menendez, A.; Esquivel, I.V. Users Acceptance of Location-Based Marketing Apps in Tourism Sector: An Exploratory Analysis. J. Spat. Organ. Dyn. 2018, 6, 258–270. [Google Scholar]
  31. Java, A.; Song, X.; Finin, T.; Tseng, B. Why we twitter. In Proceedings of the 9th WebKDD and 1st SNA-KDD Workshop on Web Mining and Social Network Analysis—WebKDD/SNA-KDD 07, San Jose, CA, USA, 12–15 August 2007. [Google Scholar]
  32. Chisholm, E.; O’Sullivan, K. Using Twitter to Explore (un)Healthy Housing: Learning from the #Characterbuildings Campaign in New Zealand. Int. J. Environ. Res. Public Health 2017, 14, 1424. [Google Scholar] [CrossRef]
  33. Neethu, M.; Rajasree, R. Sentiment Analysis in Twitter Using Machine Learning Techniques. In Proceedings of the 4th International Conference on Computing Communications and NetworkingTechnologies (ICCCNT), Tiruchengode, India, 4–6 July 2013. [Google Scholar]
  34. Aggarwal, A.; Kumar, S.; Bhargava, K.; Kumaraguru, P. The follower count fallacy. In Proceedings of the 33rd Annual ACM Symposium on Applied Computing—SAC 18, Coimbra, Portugal, 18–22 March 2018. [Google Scholar]
  35. Berg, J.A. The story of the hashtag (#): A practical theological tracing of the hashtag(#) symbol on Twitter. HTS Teol. Stud. Theol. Stud. 2014, 70. [Google Scholar] [CrossRef]
  36. Thapa, L. Spatial-Temporal Analysis of Social Media Data Related to Nepal Earthquake 2015. In Proceedings of the ISPRS—International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Prague, Czech Republic, 12–19 July 2016; Volume XLI-B2, pp. 567–571. [Google Scholar] [CrossRef]
  37. De La Pena, A. Marketing Strategies for the Social Good. Marketing 2018. [Google Scholar] [CrossRef]
  38. Palomino, M.; Taylor, T.; Göker, A.; Isaacs, J.; Warber, S. The Online Dissemination of Nature–Health Concepts: Lessons from Sentiment Analysis of Social Media Relating to “Nature-Deficit Disorder”. Int. J. Environ. Res. Public Health 2016, 13, 142. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  39. Fowler, A.; Biekart, K. Multi-Stakeholder Initiatives for Sustainable Development Goals: The Importance of Interlocutors. Public Adm. Dev. 2017, 37, 81–93. [Google Scholar] [CrossRef] [Green Version]
  40. Mellon, J.; Prosser, C. Twitter and Facebook are not representative of the general population: Political attitudes and demographics of British social media users. SAGE J. 2017. [Google Scholar] [CrossRef]
  41. Bennett, D.; Yábar, D.P.B.; Saura, J.R. University Incubators May Be Socially Valuable, but How Effective Are They? A Case Study on Business Incubators at Universities. In Entrepreneurial Universities. Innovation, Technology, and Knowledge Management; Peris-Ortiz, M., Gómez, J., Merigó-Lindahl, J., Rueda-Armengot, C., Eds.; Springer: Cham, Switzerland, 2017. [Google Scholar] [CrossRef]
  42. Tableau. 2018. Available online: https://tableau.com/es-es/node/45403 (accessed on 1 April 2018).
  43. Tlebere, T.; Scholtz, B.; Calitz, A.P. Using Social Media to Improve Environmental Awareness in Higher Education Institutions. In Information Technology in Environmental Engineering Proceedings in Business and Economics; Springer: Cham, Switzerland, 2016; pp. 101–111. [Google Scholar] [CrossRef]
  44. John, L.; Emrich, O.; Gupta, S.; Norton, M. Does “Liking” Lead to Loving? The Impact of Joining a Brand’s Social Network on Marketing Outcomes. J. Mark. Res. 2017, 54, 144–155. [Google Scholar] [CrossRef]
  45. Cao, X.; Macnaughton, P.; Deng, Z.; Yin, J.; Zhang, X.; Allen, J. Using Twitter to Better Understand the Spatiotemporal Patterns of Public Sentiment: A Case Study in Massachusetts, USA. Int. J. Environ. 2018, 15, 250. [Google Scholar] [CrossRef] [PubMed]
  46. Zhou, X.; Chen, L. Event detection over Twitter social media streams. VLDB J. 2013, 23, 381–400. [Google Scholar] [CrossRef]
  47. Saura, J.R.; Palos-Sanchez, P.; Debasa, F. El problema de la Reputación Online y Motores de Búsqueda: Derecho al Olvido. Cadernos de Dereito Actual 2018, 8, 221–229. [Google Scholar]
  48. Bifet, A.; Frank, E. Sentiment knowledge discovery in Twitter streaming data. In Proceedings of the International Conference on Discovery Science, Canberra, Australia, 6–8 October 2010. [Google Scholar]
  49. Pak, A.; Paroubek, P. Twitter as a corpus for sentiment analysis and opinion mining. In Proceedings of the LREC, Valletta, Malta, 17–23 May 2010. [Google Scholar]
  50. Honeycutt, C.; Herring, S.C. Beyond microblogging: Conversation and collaboration via Twitter. In Proceedings of the 42nd Hawaii International Conference on System Sciences, Hawaii, HI, USA, 5–8 January 2009; pp. 1–10. [Google Scholar]
  51. Kuo, T.-T.; Hung, S.-C.; Lin, W.-S.; Peng, N.; Lin, S.-D.; Lin, W.-F. Exploiting latent information to predict diffusions of novel topics on social networks. In Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics: Short Papers-Volume 2, Association for Computational Linguistics, Jeju Island, Korea, 8–14 July 2012. [Google Scholar]
  52. Boyd, D.; Golder, S.; Lotan, G. Tweet, tweet, retweet: Conversational aspects of retweeting on twitter. In Proceedings of the IEEE 43rd Hawaii International Conference on Social Systems (HICSS), Kauai, HI, USA, 5–8 January 2010. [Google Scholar]
  53. Bologna, G.; Hayashi, Y. A Rule Extraction Study from SVM on Sentiment Analysis. Big Data Cognit. Comput. 2018, 2, 6. [Google Scholar] [CrossRef]
  54. Yu, Y.; Wang, X. World cup 2014 in the Twitter world: A big data analysis of sentiments in US sports fans’ tweets. Comput. Hum. Behav. 2015, 48, 392–400. [Google Scholar] [CrossRef]
  55. Smith, A.; Brenner, J. Twitter use 2012. Pew Internet Am. Life Proj. 2012, 4, 1–12. [Google Scholar]
  56. Xu, Q.; Gel, Y.R.; Ramirez Ramirez, L.L.; Nezafati, K.; Zhang, Q.; Tsui, K.L. Forecasting influenza in Hong Kong with Google search Queries and statistical model fusion. PLoS ONE 2017, 12, e0176690. [Google Scholar] [CrossRef] [PubMed]
  57. Saura, J.R.; Palos-Sanchez, P.R.; Correia, M.B. Digital Marketing Strategies Based on the E-Business Model: Literature Review and Future Directions. In Organizational Transformation and Managing Innovation in the Fourth Industrial Revolution; IGI Global: Hershey, PA, USA, 2019; pp. 86–103. [Google Scholar]
  58. Stoové, M.A.; Pedrana, A.E. Making the most of a brave new world: Opportunities and considerations for using Twitter as a public health monitoring tool. Prev. Med. 2014, 63, 109–111. [Google Scholar] [CrossRef] [PubMed]
  59. Dolan, P.; Peasgood, T.; White, M. Do we really know what makes us happy? A review of the economic literature on the factors associated with subjective well-being. J. Econ. Psychol. 2008, 29, 94–122. [Google Scholar] [CrossRef]
  60. Li, L.; Goodchild, M.F.; Xu, B. Spatial, temporal, and socioeconomic patterns in the use of Twitter and flickr. Cartogr. Geogr. Inf. Sci. 2013, 40, 61–77. [Google Scholar] [CrossRef]
  61. Xu, C.; Wong, D.W.; Yang, C. Evaluating the “geographical awareness” of individuals: An exploratory analysis of Twitter data. Cartogr. Geogr. Inf. Sci. 2013, 40, 103–115. [Google Scholar] [CrossRef]
  62. MonkeyLearn API Reference. Available online: https://monkeylearn.com/docs/article/api-reference/ (accessed on 15 June 2017).
  63. Nvivo QSR. About Automatic Coding Techniques. 2018. Available online: http://help-nv10.qsrinternational.com/desktop/concepts/about_automatic_coding_techniques.htm (accessed on 12 August 2018).
  64. Hornikx, J.; Hendriks, B. Consumer Tweets about Brands: A Content Analysis of Sentiment Tweets about Goods and Services. J. Creat. Commun. 2015, 10, 176–185. [Google Scholar] [CrossRef]
  65. Park, S.; Griffin, A.; Gill, D. Working with words: Exploring textual analysis in medical education research. Med. Educ. 2012, 46, 372–380. [Google Scholar] [CrossRef] [PubMed]
  66. Yue, G.; Dong, Y.; Chen, H.; Lai, K. Online Textual Sentiment Analysis Technology and It’s Applications. Adv. Psychol. Sci. 2013, 21, 1711–1719. [Google Scholar] [CrossRef]
  67. Nature Geoscience. Tweets on Earth. Nat. Geosci. 2011, 4, 209. [Google Scholar] [CrossRef]
  68. Grilo, A.; Jardim-Goncalves, R. Value proposition on interoperability of BIM and collaborative working environments. Autom. Constr. 2010, 19, 522–530. [Google Scholar] [CrossRef]
  69. Herráez, B.; Bustamante, D.; Saura, J.R. Information classification on social networks. Content analysis of e-commerce companies on Twitter. Rev. Espacios 2017, 38, 16. [Google Scholar]
  70. Saura, J.R.; Palos-Sánchez, P.; Reyes-Menendez, A. Marketing a través de Aplicaciones Móviles de Turismo (M-Tourism). Un estudio exploratorio. Int. J. World Tour. 2007, 4, 45–56. [Google Scholar]
  71. Bollen, J.; Mao, H.; Zeng, X. Twitter mood predicts the stock market. J. Comput. Sci. 2011, 2, 1–8. [Google Scholar] [CrossRef] [Green Version]
  72. Saura, J.R.; Reyes-Menendez, A.; Alvarez-Alonso, C. Do online comments affect environmental management? Identifying factors related to environmental management and sustainability of hotels. Sustainability 2018, 10, 3016. [Google Scholar] [CrossRef]
  73. Hasan, A.; Moin, S.; Karim, A.; Shamshirband, S. Machine Learning-Based Sentiment Analysis for Twitter Accounts. Math. Comput. Appl. 2018, 23, 11. [Google Scholar] [CrossRef]
  74. Saura, J.R.; Reyes-Menendez, A.; Palos-Sanchez, P. Un Análisis de Sentimiento en Twitter con Machine Learning: Identificando el sentimiento sobre las ofertas de# BlackFriday. Revista Espacios 2018, 39, 16. [Google Scholar]
  75. Reyes-Menendez, A.; Saura, J.R.; Palos-Sanchez, P.P.; Alvarez-Garcia, J. Understanding User Behavioral Intention to adopt a Search Engine that promotes Sustainable Water Management. Symmetry 2018, 10, 584. [Google Scholar] [CrossRef]
  76. Saura, J.R.; Palos-Sánchez, P.; Suárez, L.M.C. Understanding the Digital Marketing Environment with KPIs and Web Analytics. Future Internet 2017, 9, 76. [Google Scholar] [CrossRef]
  77. Palos-Sanchez, P.; Saura, J.R.; Martin-Velicia, F. A study of the effects of Programmatic Advertising on users' Concerns about Privacy overtime. J. Bus. Res. 2019, 96, 61–72. [Google Scholar] [CrossRef]
Figure 1. Development of the methodological process. Source: The authors.
Figure 1. Development of the methodological process. Source: The authors.
Ijerph 15 02537 g001
Figure 2. Relationship of the nodes and number of guest reviews for environment factors identification. Source: The authors.
Figure 2. Relationship of the nodes and number of guest reviews for environment factors identification. Source: The authors.
Ijerph 15 02537 g002
Table 1. UN Sustainable Development Goals (SDGs).
Table 1. UN Sustainable Development Goals (SDGs).
GoalTitleDescription
1No povertyEnd poverty in all its forms everywhere
2Zero hungerEnd hunger, achieve food security and improved nutrition and promote sustainable agriculture
3Good health and well-beingEnsure healthy lives and promote well-being for all at all ages
4Quality educationEnsure inclusive and equitable quality education and promote lifelong learning opportunities for all
5Gender equalityAchieve gender equality and empower all women and girls
6Clean water and sanitationEnsure availability and sustainable management of water and sanitation for all
7Affordable and clean energyEnsure access to affordable, reliable, sustainable and modern energy for all
8Decent work and economic growthPromote sustained, inclusive and sustainable economic growth, full and productive employment and decent work for all
9Industry, innovation, and infrastructureBuild resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation
10Reduce inequality within and among countriesReduce inequality within and among countries
11Sustainable cities and communitiesMake cities and human settlements inclusive, safe, resilient and sustainable
12Responsible consumption and productionEnsure sustainable consumption and production patterns
13Climate actionTake urgent action to combat climate change and its impacts
14Life below waterConserve and sustainably use the oceans, seas and marine resources for sustainable development
15Life on landProtect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse
16Peace, justice, and strong institutionsPromote peaceful and inclusive societies for sustainable development, provide access to justice for all and build effective, accountable and inclusive
17Partnership for the goalsStrengthen the means of implementation and revitalize the global partnership for sustainable development
Source: United Nations [6].
Table 2. Earthquakes and corresponding hashtags.
Table 2. Earthquakes and corresponding hashtags.
Hashtag (#)LocationDateYearMagnitudeDead
#FuerzaMexicoMéxico8 September20178.2370
#KurdistanEarthquakeIran12 November20177.3630
#earthquakenepalNepal25 April20157.88964
#prayingforjapanJapan11 March20118.915,896
#HaitiHaiti12 January20107.0160,000
#ChileanearthquakeChile27 February20108.8525
Source: The authors based on Thapa [36] and Pena [37].
Table 3. Important related work.
Table 3. Important related work.
StudyAimSocial NetworkUnit of Analysis
Saura et al. [2]To identify the positive, neutral, and negative factors highlighted by visitors to Spanish hotelsTwitterPosts
Cao et al. [39]To identify behavior patterns related to time and space that explain user well-being and happinessTwitterPosts
Ekenga et al. [27]To identify the issues that most concern the community about environmental risks, by analyzing the public response on the Twitter social network to the comments regarding the high amounts of water at St. Louis of MissouriTwitterPosts
Chisholm and O’Sullivan [42]To illustrate that Twitter can provide information on housing as a public and social health problem, in a case study of the campaign #characterbuildings that started in New Zealand in 2014TwitterHashtag
Palomino et al. [36]To understand the impact of the natural environment on people’ well-being and health by identifying public opinion and studying the discourse related to nature-deficit disorder and other nature-heath conceptsTwitterHashtag
Woo et al. [43]To examine the tweets published about a natural disaster in Korea and to identify the changes in users’ sentimentTwitterPosts
Source: The authors.
Table 4. Country percentage publishing.
Table 4. Country percentage publishing.
CountryPercentage
United States7.95
United Kingdom4.64
India4.40
Nigeria3.97
South Africa3.51
Kenya3.21
Canada3.31
Spain1.99
France1.32
Australia1.32
Malaysia1.67
Mexico1.09
Belgium1.02
Source: The authors.
Table 5. Average classification of machine learning probability percentages.
Table 5. Average classification of machine learning probability percentages.
#WorldEnvironmentDayPositiveNeutralNegativeTotal
Tweets1243268719435873
Average Probability0.7190.6510.8020.724
Source: The authors.
Table 6. Sample of the Tweet classification according to the Sentiment.
Table 6. Sample of the Tweet classification according to the Sentiment.
UserTweetSentimentAverage Probability
@SeaShepherdSSCSIt’s World Environment Day! Did you know that much of the trash in our Oceans was improperly disposed of on land?Negative0.719
@VashtiHarrisonWe only have one planet, hoping we’re all doing our part to take care of itNeutral0.671
@UNYouthEnvoyHappy World Environment Day! Check out creative ideas to #BeatPlasticPollution from ppl all around the worldPositive0.930
@HammalHaidarIf someone doesn’t believe in global warming, take them to Balochistan #BalochistanTurningIntoDesertNegative0.602
Source: The authors.
Table 7. Similar Hashtags and frequency of use.
Table 7. Similar Hashtags and frequency of use.
HashtagsFrequency *
#climatechange614
#PlanetOrPlastic460
#beatplasticpollution314
#PlanetOrPlastic?229
#plastic208
#BeatPlasticPolution98
#WorldEnvironmentDay201842
#WorldOceansWeek32
#SeaLegacy32
#GetFishFarmsOut32
#Savetheplanet32
#GetFishFarmsOut23
#pollutioncrime22
#pollution20
#PlasticFreeLiving14
#worldoceansday10
#UseLessPlastic8
* Frequency = the number of times one of the hashtags was repeated within textual analysis. Source: The authors.
Table 8. Results for N1, N2, and N3 for environment factors.
Table 8. Results for N1, N2, and N3 for environment factors.
NodesCountSimilar FactorsWeighted Percentage
Node1—Positive environment factors indicators
Public Health207Diseases, Health, Healthy food, Health Equity, Safe sanitation, Essential medical products2.56
Clean Energy195Green energy, Solar energy, Biofuel, Biomass2.13
Sustainability172Energy efficiency, renewable energies, Sustainable development2.04
Node2—Neutral environment factors indicators
Sustainable cities286Smartcity, Smartcities, Green urban, green architecture2.93
Sustainable management271Energy management, Water management, resources sustainable management2.79
Innovative Ideas209Innovative challenges, sustainable ideas1.89
Promote well-being151e-Health, Healthy food, Healthy activities, Sports1.68
Node3—Negative environment factors indicators
Climate Change378Global warming, extreme weather, climate actions, climate change effects4.25
Water319Safe water, water pollution, water conservation, water resource4.02
Forests251Natural environments, Deforestation, Urban development2.98
Pollution230Climate risks, preventing pollution, acid rain, atmosphere, plastic2.75
Biodiversity146Plants, Trees, Animals, biological diversity1.84
Massive Industrialization140Healthy homes, Malnutrition, Low carbon development1.72
Source: The authors.

Share and Cite

MDPI and ACS Style

Reyes-Menendez, A.; Saura, J.R.; Alvarez-Alonso, C. Understanding #WorldEnvironmentDay User Opinions in Twitter: A Topic-Based Sentiment Analysis Approach. Int. J. Environ. Res. Public Health 2018, 15, 2537. https://0-doi-org.brum.beds.ac.uk/10.3390/ijerph15112537

AMA Style

Reyes-Menendez A, Saura JR, Alvarez-Alonso C. Understanding #WorldEnvironmentDay User Opinions in Twitter: A Topic-Based Sentiment Analysis Approach. International Journal of Environmental Research and Public Health. 2018; 15(11):2537. https://0-doi-org.brum.beds.ac.uk/10.3390/ijerph15112537

Chicago/Turabian Style

Reyes-Menendez, Ana, José Ramón Saura, and Cesar Alvarez-Alonso. 2018. "Understanding #WorldEnvironmentDay User Opinions in Twitter: A Topic-Based Sentiment Analysis Approach" International Journal of Environmental Research and Public Health 15, no. 11: 2537. https://0-doi-org.brum.beds.ac.uk/10.3390/ijerph15112537

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop