A computational methodology to measure the cultural identity of countries.

Abstract

Throughout the different waves of migrants, the adoption of the rules of the host society, as well as the propagation of their own cultures brought cultural diversity to the destination country. Although it is essential to understand societies, measuring culture and its evolution has been a complex and elusive goal, mainly due to the scarcity and the cost of obtaining data. Thus, alternative data sources are increasingly used as a data source, replacing or complementing traditional data sources. The use of social media data is an example. The growth of online social networks in recent years is impressive. Only Facebook, the most popular social network, has nearly 2.5 billion monthly active users. The growing number of Facebook users represents new potential customers from companies that pay for advertising space on the social network. In fact, most of the revenue from online social networks are concentrated on their marketing platforms. By using online social media ad platforms, an advertiser can explore micro-targeting advertising, which allows them to select users with very particular characteristics, including thousands of demographic attributes, interests, and behaviors. Social networks provide the ideal environment for inferring data from online populations, since users share a large number of personal information, as well as behavioral signals, such as likes and content shares they like. Based on this, we took advantage of the aggregated information about users provided by Facebook's advertising platform to advertisers to develop a methodology to infer the cultural identity of countries. In this work, we propose and develop a methodology to not only identify and characterize the cultural identity but also to measure cultural distance between countries based on users' interest in cultural attributes. Many cultural aspects characterize the regions in terms of cultural attributes, such as clothing, music, art, and food. Therefore, in addition to the proposed methodology, as an application example, we present a case study focused on cultural elements related to Brazilian cuisine using data from the Facebook Advertising Platform. Through this case study, it was possible to identify the typical Brazilian dishes that best relate to the cultural identity of Brazil. In addition, we measure the global spread of Brazilian food culture among countries, exploring the Facebook user's preferences for typical Brazilian dishes. The results show a high correlation between the proportion of Brazilian immigrants in each country and the distance between those countries and Brazil in terms of the proposed cultural distance. For this reason, this distance measure can complement other distance metrics applied to gravity-type models, for example, in der to explain the flow of people between countries.