A MACHINE LEARNING APPROACH TO EXPLORING SENTIMENTS OF USER DURING AN IT OUTAGE: THE CROWDSTRIKE CASE
Öz
Today, social media platforms are the fastest, most interactive public arenas, capturing the pulse of global crises and instantly shaping public sentiments. These dynamic structures multiply the speed of information spread, allowing the simultaneous interactions of millions of users to reveal different dimensions of an event. This study examined the effects of the Crowdstrike incident on July 19, 2024, a major IT outage affecting daily life, on social media platforms through sentiment analysis. A total of 1536 tweets were analysed using the Orange data mining and machine learning software. The analysis reveals that the public's reaction was multi-dimensional, with a dominant neutral sentiment. However, detailed emotion analyses based on Ekman and Plutchik algorithms showed various coexisting emotions, including joy, surprise, sadness, trust, and fear, proving emotional complexity is intertwined with information sharing and humour during a crisis. Furthermore, topic modeling analysis suggests that such a technical incident can quickly link to seemingly unrelated political agendas, highlighting social media's power to shift an event from a technical to a social and political context. These findings suggest that public response to IT crises is not merely technical but inherently social, emotional, and political in nature, underscoring the need for crisis communication strategies that account for emotional complexity and discourse dynamics on social media. This study contributes to the literature by addressing a gap in the literature through a multi-method sentiment analysis of a global IT outage, integrating VADER, Ekman, and Plutchik emotion models alongside topic modeling, and provides a replicable framework for analysing public reactions to technology-driven crises.
Anahtar Kelimeler
Sentiment analysis, Machine learning, Social media, Crowdstrike
Kaynakça
- Alghamdi, R., & Alfalqi, K. (2015). A survey of topic modeling in text mining. International Journal of Advanced Computer Science and Applications, 6(1), 147-153. https://doi.org/10.14569/IJACSA.2015.060121
- Alkaraki, S. M. S., Alias, N. B., & Maros, M. (2024). Exploring the impact of social media humor related to the COVID-19 pandemic: A systematic literature review on themes, coping mechanisms, critiques and linguistic devices. Cogent Arts and Humanities, 11(1), 2322227. https://doi.org/10.1080/23311983.2024.2322227
- Banerjee, P. (2024). CrowdStrike cyber incident vs. past major cyber incidents: Analysis and solutions. International Journal for Multidisciplinary Research, 6(4), 1-15. https://doi.org/10.36948/ijfmr.2024.v06i04.25310
- Bashir, S., Bano, S., Shueb, S., Gul, S., Mir, A. A., Ashraf, R., Shakeela, & Noor, N. (2021). Twitter chirps for Syrian people: Sentiment analysis of tweets related to Syria chemical attack. International Journal of Disaster Risk Reduction, 62(June), 102397. https://doi.org/10.1016/j.ijdrr.2021.102397
- Blei, D., Ng, A., & Jordan, M. (2003). Latent Dirichlet allocation. J Mach Learn Res, 3(Jan), 993–1022. https://doi.org/10.1162/jmlr.2003.3.4-5.993
- Bonta, V., Kumaresh, N., & Janardhan, N. (2019). A comprehensive study on lexicon based approaches for sentiment analysis. Asian Journal Computer Science Technology, 8(S2), 1–6. https://doi.org/10.51983/ajcst-2019.8.S2.2037
- Camacho, N. G. (2024). The role of AI in cybersecurity: Addressing threats in the digital age. Journal of Artificial Intelligence General Science, 3(1), 143–154. https://doi.org/10.60087/jaigs.v3i1.75
- Ekman, P. (1982). Emotion in the human face (2nd ed.). Cambridge University Press.
- El Maarouf, F. (2025). Disaster-funny in postdigital age: Memesis and the composite nature of humor in crisis. Postdigital Science and Education, 7(2), 480–499. https://doi.org/10.1007/s42438-024-00502-3
- Gabriel, R., & Röhrs, H. P. (2017). Social media: Potenziale, trends, chancen und risiken. Springer Gabler. https://doi.org/10.1007/978-3-662-53991-0