Web locales are wellspring of data for occasion location, with particular specify of the street movement action blockage and mischances or earth-quack detecting framework. In this paper, we exhibit a continuous observing framework planned for activity event identification originating from Twitter stream investigation. The framework brings tweets originating from Twitter according to a few hunt criteria; techniques tweets, by applying printed content mining strategies; to wrap things up works the order of twitter posts. The objective is to relegate reasonable class bundling to each tweet, on the grounds that related with an action of movement occasion or maybe not. The movement acknowledgment framework or system was used for continuous checking of different zones of the road arrange, considering identification of activity events only nearly in real time, frequently before on-line activity. All of us supports machine like a characterization unit, besides, we achieved an incredible precision estimation of ninety five. 75% by endeavoring a paired order issue. Every one of us were likewise skilled to segregate if movement is activated by an outside festival or not, by settling a multiclass order issue and acquiring precision worth of 88. 89%.

Keywords : Social media; Traffic detection; Text mining; Service Oriented Architecture (SOA); Twitter stream analysis.