Framing The Invisible Failure In International Press Coverage Of The 2024 Crowdstrike Crisis Through Computational Topic Modeling And Sentiment Analysis

Abstract

On 19 July 2024, a faulty update to CrowdStrike's Falcon sensor triggered a global technology outage. This study examines how the international press framed the crisis through a computational approach combining topic modeling (LDA) and sentiment analysis (VADER) over a corpus of 250 GDELT articles published between 19 and 30 July 2024. The model identified four dominant frames with asymmetric affective loads. The results reveal an inverse-scale paradox: the smallest frame concentrates the strongest negativity and causal attribution, whereas the larger frames dilute responsibility into an ambivalent neutrality. Read through Entman's (1993) four framing functions, the findings show how the media redefined a technical failure as a financial and litigation crisis, delegating corrective treatment to markets and courts rather than public regulation.