The privacy of data is one of the most essential topics in the world of privacy and security. All the advancement in technology is rendered non-essential if the data cannot be secured and be protected from various organisations. Every person or organisation desires privacy, be it technological or otherwise. It is getting more and more complicated to preserve one's privacy and maintain the confidentiality of data due to numerous policies put forth by almost every organisation to exist. Differential privacy, which makes very careful assumptions about the adversary's past knowledge, has recently become a powerful element for privacy protection. Differential privacy has received attention ever since it was initially proposed and put forth, and it is now considered to be among the most potential concepts for privacy-preserving data release and analysis in many computer science and information technology domains. In this paper, we discuss the motivation for its introduction as a tool to replace other privacy methods, use databases and datasets for the comparison of data before and after using differential privacy on it. We also explore deep into differential privacy and replicate an attack on the data by an outsider to see if the data secured by differential privacy really is protected. Comparison of the probability distributions of data before and after applying differential privacy and the variation of these distributions is also reflected in this report.

Index Terms- Differential privacy, data privacy, smart noise, data analysis, privacy preserving