Tracking Public Opinion on Turkish Political Landscape

One line of this research provides an exploratory analysis of the March 2019 Turkish local election discussions on Twitter to examine whether tracking new hashtags significantly improves the power of the model to identify the diversity of narratives of an event, as compared to tracking the generic hashtag only.The results of the analysis show that although using the generic hashtag only is as effective as using the hashtags with the highest betweenness centrality collected through the snowball sampling method, a supervised selection of the hashtags collected through the snowball sampling method appears to be more effective in collecting tweets that belong to various opinion groups, as opposed to hashtags collected through using the generic hashtag only. Therefore, supervised methods of data collection appear to be more effective as compared to unsupervised methods when the goal is to track the public opinion through online discussions.

The second line of analysis focuses on the June 2019 re-run Istanbul elections. Following a mixed methods approach, this research identifies the themes in the discursive fields and filters a large list of tweets into thematic categories using quantitative computational methods, which allows a systematic and time-efficient categorization for in-depth qualitative reading of the large text corpora.

Meltem Odabaş
Meltem Odabaş

Computational Social Scientist. My research interests include digital technology, social media use, and the relationship between social interaction and cultural formation.