Towards sentiment-driven maps showing touristic attractiveness

User generated texts on tourism-related social network sites do not only contain factual information, but also valuable opinions and ratings of locations. Nevertheless, most maps on these sites only show markers where something described in a user generated text is located. In particular, no further information is derived from the text and displayed on the maps. Moreover, generalization operations are not employed, although in most cases aggregation and displacement of the user generated content would be necessary to achieve more readable maps. Therefore, we propose a method which automatically creates user-sentiment enriched maps. We use natural language processing tools in order to mine user sentiments for specific places from user generated texts and we propose specially designed point symbols which represent the corresponding mined user sentiment for each location. Additionally, we propose a heuristic, based on Voronoi diagrams, which slightly displaces the aforementioned symbols in case they are very close. This makes the provided map easier to read.