Social media and online services with user-posted content have generated a staggering amount of information that has potential applications in various areas such as incident analysis and emergency management. In this research, the authors explored the feasibility of using social media data for detecting traffic incidents and collecting supplemental incident information. A comprehensive approach has been developed to extract and analyze real-time traffic related twitter data for incident management purpose. The developed approach consists of three steps: (1) Development of traffic incident related key words and their association rules; (2) Extraction of real-time tweets with influential word sets; and (3) Ranking and selection of traffic related tweets. The developed approach was implemented at District of Columbia Department of Transportation for incident management. Data validation has been conducted against the real-world incident database. The preliminary results of analysis have shown that social media data are promising for early incident detection and can be used as supplemental source for incident data collection.