Social media platforms have become a breeding ground for various criminal activities, leading to the need for a comprehensive survey to understand their nature and detection methods. The vast amount of data available on these platforms requires advanced analytical techniques to derive meaningful insights. In this paper, we present valuable insights from multiple studies, addressing crimes, event detection, and machine learning applications on social media. We propose methodologies for crime detection, event detection, and the utilization of machine learning algorithms, emphasizing the significance of interchangeability and efficiency in this context. Our findings suggest that advanced analytical techniques can help improve crime detection and event extraction, and machine learning algorithms can be effectively utilized to achieve this goal.