George Floyd Four Years After: A Data-Driven Analysis of Posts and Comments on X (Formerly Twitter) - Acta Universitatis Sapientiae

Acta Univ. Sapientiae, Social Analysis, 14 (2024) 16–32

DOI: 10.47745/aussoc-2024-0002

Abstract. The death of George Floyd in the hands of the Minneapolis Police on 25 May 2020 led to public outcry, followed by a worldwide protest against the rampant killing and humiliation of black people by the police in the western hemisphere, especially in the United States.The objectives of this study are to use data mining techniques and machine learning algorithms to better understand how the online communications emanating from X (formerly Twitter) trended during the period of the protests, and the observed characteristics of these communications.Due to the large volume of data collected from the social media platform X, two separate datasets in the form of posts (formerly known as tweets) were collected in DataFrame format using the Twitter Archival Google Sheets (TAGS). The first dataset was collected using #BlackLivesMatter and the second using #GeorgeFloyd.Using modules from the Python Pandas ecosystem specifically designed for data analytics, operations such as sentiment analysis, word count, and data visualizations such as word cloud were made possible.The social network package Gephi was found most suitable for analysing the network that evolved over the period under review.Our social media analytics of the #BlackLivesMatter dataset showed that 40% of the tweets analysed were positive, 44% were found to be neutral, and only 21% were categorized as negative by the TextBlob algorithm.A simple network was observed to have evolved due to the proximity in location of social media handles.Using the #GeorgeFloyd dataset, our analysis showed that 39% of the tweets were positive, another 39% were found to be neutral, and only 22% were considered negative by the algorithm for sentiment analysis this time around.Overall, the comments on Twitter were found to be positive and in support of the protests and clamour for change, social justice, police reforms, equality, and equity.

Keywords: BlackLivesMatter, George Floyd, social media, social network analysis, sentiment analysis, social media movements, Twitter, tweets