Temporal Data Analysis for Event Detection in Social Media
Loading...
Authors
Aldhaheri, Abdulrahman
Issue Date
2015-03-27
Type
Presentation
Language
en_US
Keywords
Data analysis , Statistial tool , Social media , Event detection
Alternative Title
Abstract
This research examines the possibility of detecting events using very basic statistical tools. We perform our analysis on a sample dataset collected from Twitter. We show that due to the large amount of data generated by active twitter users, simple statistical tools could be use to detect patterns that represent social events. We have collected posts generated by users over three weeks time period starting from Sep. 17 until Nov. 20. We have filtered these data by collecting only ’NFL’ related posts using twitter API. The amount of tweets was more than 4.4 million tweets. The size of the dataset was more than 17 Gb. We show in our analysis that events could be detected using basic social media network analysis. We’ve been able to successfully detect events by only using volume and frequency analysis.