Mapping the Urban Unconscious: Towards a Hybrid Application of Deep Mapping and Sentiment Analysis
Guilherme Giolo | Lecturer in Sociology, Erasmus University Rotterdam, Netherlands
Yorgos Paschos | PhD Researcher, University of York
Digital place-making has long been associated with web-based and neogeographical representations of place while digital deep mapping has been utilized as a tool for the creation of place-making practices. It is a sociotechnology which aims to generate multidimensional descriptions of places that include historical, political, economic, cultural and social information extracted from archives and personal stories. Hence, digital deep mapping combines maps, statistics, newspaper articles as well as personal stories such as memoirs, interviews and conversations while also enabling locals to create a sense of place by taking photographs and commenting. It becomes a collaborative projection of place-making practice in which community-based participatory representations of places merge and interact.
Considering the textual form of the personal stories, biographies, newspaper archives that go into the construction of digital deep maps, it is possible to conceive of a sentiment analysis approach to deep mapping practices that incorporates all this textual material. Sentiment analysis is a research method characterized by examining the positive or negative tone of a text. Since sentiment analysis filters a large number of documents, it is a useful tool for understanding mood changes and other psychological characteristics underlying many texts. We propose that a combination of digital deep mapping with sentiment analysis provide a useful way to understand how the sentiment of specific communities evolves during time. In such a way, it is possible to understand how locals’ quality of living and placemaking dynamics change over time, and in what terms this change can be perceived.
Such a methodological operation consists of firstly constructing a deep map of a specific location. Here, we propose the example of Amsterdam Noord, a neighborhood of Amsterdam undergoing gentrification, after decades serving as a shipyard and industrial park. The next step consists of individuating the textual material of the deep map which will serve as a corpus for the sentiment analysis. Lastly, the final research phase is conducting a sentiment analysis on the selected material. The significance of this approach lies in relating the observed changes in sentiment throughout time to external events, in order to see if and how local sentiment changes. This combination will result in more visualizable descriptions such as a map of the region to which the textual material refers.