This paper intends to create an application that performs the Data Mining with textual information linked to geolocation data. The structure of the information is distributed in heterogeneous and complex scenario that presents a Social Network. The purpose is that on the final extraction of information, the results are worked for Social Recommendation. However, the recommender systems present some failures in the filtering of the results and the way they are suggested to users. Then, the article presents a methodology of social recommendation to Location-Based Social Network with text mining techniques, and expose issues that still need to research more effective and consolidated results.