This paper presents a big data analytics system that allows government to measure citizens' perception of security from the Twitter social network. The proposed system was built using a set of big data tools in order to collect, pre-process, classify, index, and visualize data. The system is able to detect whether a tweet is related to security, which is used to present in a heat map the perception of security of a city. A machine learning algorithm was trained to learn to recognize security characteristics in tweets. Results show that this system is a powerful tool to visually analyze how citizens perceive security.