Problem Statement

Information has always been key to making informed decisions. With the emergence of a considerable number of disciplines requiring insight from data in order to enable them provide better services or simply for the aim of contributing to information and knowledge for research and consumption purposes, there have been need to come up with diverse sources of such data (information). However, accessing reliable and credible information has proven to be a challenge over years with a large number of available information, either being expensive to obtain or biased or even both.

In our project, we will explore twitter as a data source to analyze and report on the social problem of “The role of politics and political parties in a society’s cohesion”. Through this, we will proof or deny the credibility of social media as an upcoming and recent digital revolution

Our project intends to cover

1. General concepts about the Australian politics
2. The state of polarization of the Australian voting body

• Textual data mining

1. Basic data analysis and visualization
2. Network analysis
3. Effects of politics on day-to-day society relationships

Our project may not cover:

1. Extensive research on the scope of Australian politics
2. In-depth data analysis and data mining

• Background relationship of politics and polarization

Politics as we know it is an old art, stretching back to thousands of years if not the dawn of civilization. Maitah (2007) defines politics as, “ the activity through which people make, preserve and amend the general rules under which they live.” further Maitah (2007) argues that politics can be viewed as competition among individuals or even groups chasing their interests. More often, politics is characterized with a number of ideologies which are compelling enough to attract member who share the common vision, held together by their belief in the same and importance of their ideology as being able to push for their interests and that of others if adopted. The role of politics is therefore intricately interwoven with the society such that it makes up an important aspect of order and management.

Diversity in human nature dictates difference in opinions and beliefs therefore breeding conflict among members of a society in the choice of ideologies to follow. More often, this beliefs breed better ideas on leadership when taken up positively. However, recent developments of politics has led to complex adoption of ideologies leading complications such as polarization. In a study of the American society on how members of a different political party view those of different political parties, Fingerhut(2018) notes that “the top reason to belong to a party is a belief that its policies will benefit the country” however she argues that a big majority, “cite the other party’s harmful policies as a major factor, according to a new national survey.” (Fingerhut, 2018).

Research Objectives

In this research project we will explore the effects of political affiliation on inter-individual relationships as well as intra-individual relationships using an Australian case study of three political parties and data drawn from twitter handles of the political leaders and their friends.

Our study aims to establish existence of a relationship between politics and society, second, we aim to come up with answers to questions such as:

1. Are political parties disjoint?
2. Do members of different political parties hate each other?
3. Do members of different political parties maintain relationships with those are not?

Previous studies undertaken independently have proven existence of polarization in Australia ,upon completion of our project we will be able to contribute to the growing debate of claims of political polarization and decide on the truths of the claims. We will then suggest further studies on a topic depending on the outcomes of our research, most important we will determine the distribution of the Australian voters.

One of the motivations for undertaking this project was to be able to understand the length with which politics and political ideologies have encroached into our lives

Our research methodology is divided into:

• Data exploration which involve

1. Data mining
2. Data cleaning

• Data analysis
• Hypothesis description
• Credibility of data
• Project assumptions

To obtain data for this project we mine text data from social media using an R package “twitteR” from twitter handles of 3 politicians and their friends whom we assume to be members of their political parties and import it to R for analysis.

We explored the imported data for outliers and missing values after which we examine the distribution of the data.

The process of data analysis involved exploring for relationship among the data, determining the connection of the original data with a larger set of data, creating a data network, calculating density and neighborhood of the data edges, examining existence of Homophily among two groups of political parties and lastly determining the nature of the network we create.

In order to answer a number of research questions we set out three hypotheses, one null hypothesis and two alternative hypothesis:

H_0: Australian political parties are disjoint

H₁: Australian political parties are not disjoint

H₂:Members of different political parties are not friendly to each other

For our project the initial sample is of size (n=30) and the second sample is of size (n=30000) statistically our second data sample size is credible for use to provide assumptions on the population. Social media is currently one of the avenues to explore the thoughts of users on current affairs as well as obtain insight of the relationship of users with each other. Therefore our data source is fairly suitable to mine data for our analysis.

1. All the twitter friends of the three politicians are members of their political parties
2. The twitter friends of the friends of politicians have a positive relationship with each other, I.e. they are not enemies despite either being of the same political ideology or holding a different political ideology

• The political landscape of Australia is uniformly distributed such that one is either a member of one given political party or the other

Methodology

From our research analysis results, we note that, there are no shared friends among the three politicians. Partially it may be due to the assumption of unwavering loyalty to a given politician or probably due to the limit of our sample size. However, assuming that the sample size is sufficient enough to predict the trend in the population, we can suppose that each member or friend of a politician does not support  any other party or ideology apart from theirs. In addition, the scenario of singular party support may be assumed to be due to polarization. However, our study was limited in the sense that we only explored relationships and not the reason behind the kind of relationship.

 Despite there not being a friend who supports two party leaders at the same time, from our research we identify an existence of a relationship between the members. I.e. there were there were 26 friends shared by the friends of the politicians. It is with that data that we build a network to examine both the kind and structure of the relationships

According to Garson (2006) social capital broadly refers to , “resources inherent in social relations which facilitate collective action…” Therefore social capital include:

1. Trust
2. Norms

• Networks of association among social groups

In our research, the measure of social capital is inclined toward relationship among the groups of interest, from the network graph we denote that there are no nodes exhibiting higher social capital, i.e. the social networks are relatively even distributed. We therefore assume that there is a decline of the social norm of reciprocity, which Garson attributes to encourage “…bargaining, compromise, and pluralistic politics…”

Our data network has a low density of 0.01693989 which signify little connections among the nodes

Conclusively, the average neighborhood of between 2 and 3 for all the nodes and low density imply little connection between the groups

From our graph of Homophily we note that, there is some connection from the labor party to the other parties. This may be due to a range of factors which may include some similar ideology that push for considering either of the parties, however the low percentage of Homophily indicate disjointedness among the political parties. Cook et al. (2006). In their paper on Homophily in social networks argue that similarity breeds connection, they call it the “Homophily principle”

Ravasz and Barabási (2003), proposed that “real networks are scale free and transitive in nature which can be seen as a consequence of the hierarchy underlying the network…” as such, in our hierarchy where it was scale free and transitive, hence a real network. Helic D. and Strohmaier (2013) explored the usefulness of tag hierarchies. They argue that “nodes with high degree centrality are connected to a small number of high and mid degree nodes (high centrality) as well as to a large number of low degree nodes (low centrality).” Therefore following their argument and given that very few nodes of our network are connected to the higher branches, we conclude that our network has high centrality and hence weak.

Conclusion

To answer our research questions we draw inferences from our data analysis and discussion. From which we noted that:

1. There was disjoint among the political parties and we therefore adopt/accept the null hypothesis and reject the first alternative hypothesis
2. There was a relationship among members of different political parties indicating that despite political differences, there was still positive relationships and hence the assumption that difference in political affiliations and ideologies leads to hate is null.

In conclusion, politics has always been here and always will be. Our adoption and way of dealing with the contemporary issues accompanying its daily evolution will be the defining force of our fabric as a society. Polarization has proven to negatively affect the society and ways in which to mitigate it ought to be sought.

Given this research project, we would like to suggest further research on the contemporary issue of polarization and ways in which to curb it.

References

Gilbert, E. and Karahalios, K. (2009) Predicting Tie Strength with Social Media. In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, ACM, New York, 211-220.

Girvan, M. and Newman, M.E. (2002) .Community Structure in Social and Biological Networks. Proceedings of the National Academy of Sciences, 99, 7821-7826.
Available from: https://doi.org/10.1073/pnas.122653799

Gibson, R., & Cantijoch, M. (2013). Conceptualizing and measuring participation in the age of the Internet: Is online political engagement really different to offline? Journal of Politics, 75(3), 701–716.[Crossref], [Web of Science ®], [Google

Scholar]

Li, X., Chen, H. & Li, S.(2010). Exploiting emotions in social interactions to detect online social communities. In Pacific Asia Conference on Information Systems,pp 136.

Best, S. J., & Krueger, B. S. (2005). Analyzing the representativeness of Internet political participation. Political Behavior, 27(2), 183–216. [Crossref], [Web of

Science ®] ,[Google Scholar]

Yang, B, Cheung, W, and Liu,J. (2007). Community mining from signed social  networks.IEEE Trans.Journal on knowledge and data engineering, Vol 19(10):  pp1333–1348,

Larusso, N., Bogdanov, P. & Singh, A. (2010). Identifying communities with coherent and opposing views. In Proceeding of the 15th Annual Graduate Student Workshop in Computing. Santa Barbara : UCSB, pp 31–32 .

Putnam, R. D. (2005b). Tuning in, tuning out: The strange disappearance of social capital in America. Political Science and Politics 28: pp 664-683.

Davis B, Aldrich H. 2000. The organizational advantage? Social capital, gender and  small business owners’ access to resources.Pap.pres Am.Social. Assoc. mtg., Washington, DC.