27 Mar 2024
09:00 Doctoral defense IC3 Auditorium
Theme
Extracting Urban Perceptions from Texts Shared on Social Media: Framework and Applications
Student
Frances Albert Santos
Advisor / Teacher
Leandro Aparecido Villas - Co-supervisor: Thiago Henrique Silva
Brief summary
The automatic extraction of urban perception shared by people in location-based social networks (LBSNs) is an interesting multidisciplinary research objective. Urban perception facilitates the understanding of the intrinsic characteristics of urban areas in a scalable way, helping to leverage new services. However, the content shared in LBSNs is diverse, in almost every way, making the task of extracting urban insights challenging. Considering the messages shared in LBSNs, we propose a generic framework to support the extraction of urban perception from text messages written in natural language. We demonstrate the potential of our framework using an approach that
explores spatio-temporal and semantically similar opinions in the context of urban outdoor areas. Studying Chicago, New York and London, we show that LBSN data provides valuable information about urban areas. To analyze and validate our results, we performed a spatiotemporal analysis to measure concentration and stability of perception over time. We show that our framework can be used to extract useful urban insights, helping to better understand urban areas in many aspects, such as the feeling reflected by neighborhoods and keywords used by people to qualify their experiences, which make up our outdoor urban perception dictionary, called UOP-dictionary. Furthermore, we conducted a comparative analysis based on a public dataset, which contains volunteers' opinions about urban areas expressed in a controlled experiment. We observe that both results produce a very similar level of agreement. Finally, we developed a demonstration that aims to improve the real estate market, with the aim of demonstrating the benefits of a better understanding of urban perceptions and, thus, being able to provide smart services to citizens.
Examination Board
Headlines:
| Leandro Aparecido Villas | IC / UNICAMP |
| Julio Cesar dos Reis | IC / UNICAMP |
| Nádia Félix Felipe da Silva | INF/UFG |
| Vinícius Fernandes Soares Mota | INF/UFES |
| Fabio Augusto Faria | ICT / UNIFESP |
Substitutes:
| Sandra Eliza Fontes de Avila | IC / UNICAMP |
| Hélio Pedrini | IC / UNICAMP |
| Roberto Alencar Lotufo | FEEC / UNICAMP |
| Heitor Soares Ramos Filho | DCC / UFMG |
| Helder May Nunes da Silva Oliveira | CECS / UFABC |
| Humberto Torres Marques Neto | PUC Minas |