Google Drive: link
Independent research: How the consensus is built and involved through a voting mechanism? Using Taiwan’s voting regarding Uber compliance issue on Polis as an example
Polls and probabilistic modeling results often serve as a reference of how people voting behavior in the final election. However, ‘horse race’ problem and misinterpretation of the model can distort the original intention of polls and voting, which should be a process to unite society and find common ground among different communities. Also, along with the rising of modern social media, echo chambers can be formed and intensify the polarization process. Therefore, rethinking the voting process and reorienting it to a mechanism is essential to embark on healing. This paper would discuss how the voting mechanism can be used to build consensus and lift the echo chamber. It will first plot time-series data to investigate the evolution of consensus on the opinion spectrum, before applying the UMAPs to explore the high-dimension characteristics and calculate the social distance, including the suggestion on the comment displaying algorithm. Also apply the NLP’s sentiment analysis technique to classify the emotion in the comment data.
Polis is a platform, where everyone can draft a statement about how a matter should be solved and respond to other participant’s suggestions by either agreeing or disagreeing with them. According to the individual response, Polis churned through the many axes of agreements and disagreements to draw a map (UMAP), then it will reduce the high dimension to the four dimension space and try to show people different opinions from different groups.
- For Participants' vote data participants-votes.csv
- For comments data comments.csv
The research may encounter the problem of interpretation of variables in the dataset, and self-learning of algorithms: such as PCA and Leiden graphs, UMAP, etc.
The research may encounter the problem of interpretation of variables in the dataset, and self-learning of algorithms: such as PCA and Leiden graphs, UMAP, etc.