Deployment of Phasor Measurement Units (PMU) in the United States transmission grid has brought a new data stream to be processed and an opportunity to improve situational awareness on the grid. This new data stream offers opportunity for a faster detection and response algorithm to minimize wide spread outages. High rate of data collection of PMU systems has also brought a challenge on how to extract information from fast moving PMU data stream in real time to improve situational awareness inside a control room. Despite the fact that mathematical and probabilistic methods are the most accurate methods of stability analysis, online decision making algorithms cannot afford the latency brought by those methods. Traditional batch processing Artificial Intelligence (AI) techniques have been extensively studied as potential replacements for these approaches, however conventional AI techniques do not deal with continuous streams of fast moving phasor data. This paper presented a novel application of the stream mining algorithms for synchrophasor data to meet quick decision making requirement of future situational awareness applications in power systems. To prove that the proposed methods are efficient and capable of handling huge amounts of data with reasonable accuracy and within limited resources of memory and computational power, four different experiments with different conditions (changing/unchanging the load conditions of Real Power and Reactive Power, fixing the size of memory, and comparing the performance of non-adaptive Hoeffding tree with traditional decision tree algorithms) were conducted. The algorithms discussed in this paper support decisions inside the control rooms helping stakeholders make informed decisions to improve reliability of the future smart grid.