Reconstruction and analysis of twitter conversation graphs
Proceedings of the First ACM International Workshop on Hot Topics on …, 2012•dl.acm.org
User interactions over social networks has been an emergent theme over the last several
years. In contrast to previous work we focus on characterizing user communications patterns
around an initial post, or conversation root. Specifically, we focus on how other users
respond to these roots and how the complete conversation initiated by this root evolves over
time. For this purpose we focus our investigation on Twitter, the biggest micro-blogging
social network. To the best of our knowledge this is the first such method that is able to …
years. In contrast to previous work we focus on characterizing user communications patterns
around an initial post, or conversation root. Specifically, we focus on how other users
respond to these roots and how the complete conversation initiated by this root evolves over
time. For this purpose we focus our investigation on Twitter, the biggest micro-blogging
social network. To the best of our knowledge this is the first such method that is able to …
User interactions over social networks has been an emergent theme over the last several years. In contrast to previous work we focus on characterizing user communications patterns around an initial post, or conversation root. Specifically, we focus on how other users respond to these roots and how the complete conversation initiated by this root evolves over time. For this purpose we focus our investigation on Twitter, the biggest micro-blogging social network. To the best of our knowledge this is the first such method that is able to reconstruct complete conversations around initial tweets. We propose a robust approach for reconstructing complete conversations and compare the resulting graph structures against those obtained from previous crawling strategies based on keyword searches. Our crawl provides a large scale dataset, ideal for computer scientists to run large scale experimental evaluations, however our dataset is made of a collection of small scale, highly controlled and complete conversation graphs ideal for a sociological investigation. We believe our work will provide the proper dataset to establish concrete collaborations with interdisciplinary expertise.