Detection and Analysis of Twitter Propaganda on Russia-Ukraine War

Ethiopia Gebreselassie

Master's thesis

Abstract:

In this digital age, X (Formerly known Twitter) has transformed to become into more that just a social media platform; It has become a key player in the course Russia-Ukraine conflict, where 280 characters can carry the weight of war. This thesis presents a technical exploration into the detection and analysis of X (Formerly known twitter) bot detection on the tweets concerning the Russia-Ukraine conflict. This study used an X (Formerly known twitter) dataset collected in the period of December 2021 to March 2022. This data was obtained from one of the well-known sites, Kaggle. Initial analysis involved descriptive statistics on different columns of the dataset, which provided a general understanding of the data distribution and patterns. The study then progressed into text analysis, extracting textual content to understand the language. A suite of machine algorithms, including Isolation Forest, K-means, and one-class SVM, were employed to identify patterns in the tweets, detect anomalies, and classify instances, assessing each algorithm's effectiveness in detecting anomalies and meeting the research objectives. Ultimately, this study is a map of modern storytelling in a time of conflict. It serves as a guide for future explorers and a call to action for collaborative guardianship of truth in our interconnected world. This work stands as a testament to the enduring power of authenticity and the shared responsibility to maintain it amidst the ever-changing currents of online dialogue.

Keywords

bots, tweets, search words, Russia, Ukraine, Conflict

Language used: English

Date on which the thesis was submitted / produced: 2. 4. 2024

Thesis defence

• Supervisor: doc. Ing. Miloš Ulman, Ph.D.
• Reader: Dino Sodamin, externi

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme:
Informatics