Twitter has been an excellent tool for extracting information in real-time, such as news and events. This dissertation uses Twitter to address the problem of real-time new events detection on the Storm distributed platform such that the system benefits from the scalability, efficiency and robustness this framework can offer. Towards this direction, three different implementations have been deployed, each of which having a different configuration. The first and simplest distributed implementation was the baseline approach. Two implementations followed, in an attempt to achieve faster data processing without loss in accuracy. The rest two implementations demonstrated significant improvements in both performance and scalability. Specifically, they achieved a 1357.16 % and 1213.15 % speed-up over the single-threaded baseline version, correspondingly. Moreover, the accuracy and robustness of the scalable approaches comparing to the baseline version were retained. Using Storm for Real-Time First Story Detection