The use of cryptocurrencies has grown significantly in recent years, and so has the number of fraudulent activities involving them. In this study, our objective is to identify suspicious patterns in cryptocurrency transactions using exploratory data analysis (EDA) and machine learning techniques. We analyzed a dataset of transactions from both suspicious and random addresses and discovered several patterns indicative of suspicious activity, suggesting that certain accounts function as hubs for illegal activities. Our findings validate speculated commentaries in cryptocurrency markets and contribute to the growing body of research on using machine learning in fraud detection. We highlight the importance of advanced analytics in identifying and preventing financial crimes. Our framework provides actionable insights for regulators and financial institutions to enhance cryptocurrency ecosystem security and integrity.