By analyzing data from the London unrest in 2011, researchers at Cardiff University pointed out that computer systems can scan Twitter's content and find potentially dangerous events, such as glass destruction or burning cars, much faster than waiting to be reported to the Metropolitan Police Service.
The system can also use rumors on Twitter to estimate where the unrest occurred and even track the position of the community gathered in real-time, as mentioned by Engadget.
The researchers used a set of machine learning algorithms to analyze each tweet from a dataset of 1.6 million tweets. In addition to tweet content, this algorithm also takes into account when a tweet is created and the location of the uploader. With two exceptions, this algorithm can always track events faster than the police.
The study came after West Midlands Police Chief Police said, due to budget cuts, the police would "face serious challenges" if it had to deal with unrest as of 2011. Researchers from this report believe monitoring Twitter can answer the problem.
Dr Pete Burnap from the School of Computer Science and Informatics, Cardiff University said, "Prior to this, we have used machine learning and natural language processing in Twitter data to better understand online irregularities, such as the spread of antagonistic content and cyber hatred."