Show simple item record

dc.rights.licenseCC-BY-NC-ND
dc.contributor.advisorZeng, Jing
dc.contributor.authorTalu, Uygar
dc.date.accessioned2023-10-12T23:01:03Z
dc.date.available2023-10-12T23:01:03Z
dc.date.issued2023
dc.identifier.urihttps://studenttheses.uu.nl/handle/20.500.12932/45356
dc.description.abstract[""This study investigates the digital presence of the Justice and Development Party (AKP) and the Republican People's Party (CHP) on TikTok. It aims to elucidate their political stance and digital marketing strategies using machine learning techniques. The research combines data acquisition and advanced analytical processes to decipher the political narratives and strategies of each party. Data was categorized into profiles and hashtags related to AKP, CHP, and general discussions. Unique TikTok links streamlined the comprehensive data extraction process, with Python-based libraries assisting in retrieving engagement metrics and video descriptions. The choice to analyze the 'general discussion' category was determined by the study's progression and focus. Several approaches were applied in the Machine Learning Analysis phase. The Isolation Forest algorithm was essential for its computational efficiency, highlighting videos with distinct engagement patterns. Face Detection and Emotion Recognition assessed the implicit emotions in the content, playing a crucial role in distinguishing the content and strategy between parties. Speech Recognition transcribed spoken elements from videos, paving the way for Topic Modeling to determine the central themes the parties addressed. The culmination of the analysis used the K-means algorithm. This clustering aimed to highlight competition in content creation, offering a nuanced comparison based on the narratives each party emphasized. It also facilitated a holistic strategy analysis, enabling parties to gauge and respond to competing narratives. Results show the effectiveness of machine learning in unveiling political strategies on TikTok. Face detection, Emotion detection, topic modeling, and Clustering analysis elucidate the tactics each party adopts to maximize interaction and spread their ideologies. The research underscores the significance of digital platforms in modern political discourse, offering a fresh perspective on social media's potential in understanding political strategies""]
dc.description.sponsorshipUtrecht University
dc.language.isoEN
dc.subjectThe subject of the thesis is the analysis of the digital presence, political stance, and digital marketing strategies of the Justice and Development Party (AKP) and the Republican People's Party (CHP) on TikTok using machine learning techniques.
dc.titleDeep Dive into Political Communications on TikTok: Insights from Sentiment Analysis, Computer Vision, Natural Language Processing in the context of Turkey 2023 Elections
dc.type.contentMaster Thesis
dc.rights.accessrightsOpen Access
dc.subject.courseuuApplied Data Science
dc.thesis.id25153


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record