The spatial and textual properties of Points of Interests (POI) provided by different Location-Based Services (LBS) may be inconsistent, inaccurate or may be missing from one or several providers. The same query executed with several tourist LBS providers may generate different and inconsistent results for spatial and textual properties. Moreover, graphic symbols of legends are also different from one provider to another. The integration process of these heterogeneous spatial data, in a dynamic and scalable context, using incomplete sources with variable quality of data, is a challenging issue. In this Ph.D., this challenge is addressed to integrate POIs and to represent them.