Page 850 - Proceedings Collega2023

collected from the first time it was posted by KiniTV on facebook on 17 of March 2023 until 23 July 2023,
consists of 1087 comments. To maintain ethical standards and respect user privacy, all data collected are
publicly accessible and do not involve any unauthorized access to private information. The collected data
undergo a rigorous preprocessing phase, which includes text cleaning, removal of duplicates, and
anonymization of user information. Textual content is tokenized, and any personally identifiable
information is redacted to ensure anonymity and compliance with ethical guidelines.

Correlational Analysis

This approach employed to identify relationships and associations among the various themes
that emerge from the ethnographic data. This statistical technique allows us to uncover patterns of co-
occurrence, causality, and interdependence among different themes, providing insights into the
underlying structure of political discourse. To run this complex analysis, I am using RStudio with several
main packages such as iGraph and ggraph. By combining virtual ethnography with correlational analysis,
I seek to transcend the limitations of traditional sentiment analysis methods and gain a deeper
understanding of the key themes that shape online political discussions. This research methodology
allows us to explore the multifaceted and dynamic nature of social contentious discourse, providing
valuable insights for academia, and society as a whole.

Analysis and Result

The analysis has been conducted in several different segments, including an overall comment, as
well as categorically by ethnic groups such as the Malay, Chinese, Indian, and unknown ethnic
backgrounds. The analysis consisted of two measures. The differences are in the number of words that
appear or used in interactions have been categorized into two criteria: those used more than 5 times and
those used more than 10 times, which split the measure into two part, the general themes and also
i
primary theme. For some groups, the number of comments are a bit low compare to the other.
For instance, the comments from Indian ethnic group in this matter is not many as Malay and
Chinese, hence the then number of words that appears are also lowered in the analysis to ensure that
the meaning and theme of the interactions are successfully translated. From the words networks that
appeared, we only select word combinations that are relevant and meaningful to the discussion. The
analysis procedure is divided into three main steps. Firstly, it involves identifying and understanding the
emerging themes. Secondly, it deems it important to consider uniformity and balance within the word
combinations' interactions in order to understand them. At the third stage, the meaning is assessed
based on the emerging themes, followed by discussions on the findings.
The diagram in Figure 1 illustrates the general themes that occurred within the entire sample taken,
where the correlation was at 3% or higher, and n≥5. From this interaction, we can clearly observe several
words networks that appear to form the overall interaction themes. Figure 2 presented the primary
themes at 3% or higher, and n≥10.

Figure 1 Overall Comments n ≥ 5

International Conference on Local Wisdom of the Malay Archipelago (COLLEGA 2023) Page - 837 -