A large number of companies, organizations and other entities collect and elaborate personal data from people, which are frequently published for research or other promotional purposes. This paper deals with the effective anonymization, in applications that store data in relational databases. The optimum choice of a privacy model along with its application can the effectively protect personal data and allow low percentage of information loss. In this context and, as far as the need of anonymization of a relational database is concerned, a tool, named CRPM, has been developed, which is able to propose a suitable privacy model after the data recipient has viewed and answered questions related to the dataset to protect the disclosure of sensitive data.