Databases contain potentially valuable knowledge that should be processed and interpreted in order to be useful. In this context, we examine synergy of fuzzy functional dependencies and linguistic summaries. Strength of dependency between two examined attributes in the whole database can be detected and expressed by fuzzy functional dependencies. If dependency exists but is not sufficiently strong, linguistic summaries are a convenient method for revealing the character of dependency. Both methods simulate human reasoning in looking for the knowledge by expressions of linguistic terms which ensure that similar entities are treated similarly and mined knowledge is represented in an understandable way. The database of municipal statistics is a suitable for mining relational knowledge due to a variety of collected data which could provide significant support for policy and decision making. Finally, we conclude by discussing practical issues and topics for further research.