Criação de assinaturas culturais de áreas urbanas usando dados geolocalizados da WEB

Abstract

Research concerning the study of the different cultures that exist in the world is carried out, in large part, in traditional ways, which are expensive and do not scale. As an alternative, several recent studies benefit from web data. In this case, it is also attractive due to the need to achieve results from time to time, as culture is a changing entity. Having knowledge of the characteristics of different cultures and identifying cultural similarities between areas can bring several economic and social benefits, such as recommending places according to cultural criteria. However, the concept of culture is quite complex, making it challenging to find data encompassing its different aspects, such as religion, cuisine, and arts. Thus, this work aims to obtain relevant characteristics of urban areas, using geolocated data from web sources, but enlarging the dimensions to allow the exploration of the greatest possible number of cultural aspects. Then, creating digital signatures that would enable identifying areas with cultural similarities. The results generated from a Hierarchical Clustering technique are very consistent and even surprising, such as the separation of Curitiba’s neighborhoods into culturally characteristic clusters and the segregation of cities that culturally resemble cities from other continents.