Studying urban space from textual data: Toward a meth-odological protocol to extract geographic knowledge from real estate ads. - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2022

Studying urban space from textual data: Toward a meth-odological protocol to extract geographic knowledge from real estate ads.

Résumé

Real estate ads are a rich source of information when studying social representation of residential space. However, extracting knowledge from them poses some methodological challenges namely in terms its spatial content. The use of techniques from artificial intelligence to find and extract knowledge and relationships from textual data improves the classical approaches of Natural Lan-guage Processing (NLP). This paper will first conceptualize what kind of infor-mation on urban space can be targeted in real estate ads. It will then propose an automated protocol based on artificial intelligence to extract named entities and relationships among them. The extracted information will finally be modeled as RDF graphs and queried through GeoSPARQL. First results will be proposed from the case study of real estate ads on the French Riviera, with a focus on toponymy. Perspectives of quantitative spatial analysis of the geolocated RDF models of real-estate ads will also be highlighted.
Fichier principal
Vignette du fichier
Blanchi_Fusco_Emsellem_Cadorel_CAMERAREADY.pdf (1.08 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03925785 , version 1 (05-01-2023)

Identifiants

Citer

Alicia Blanchi, Giovanni Fusco, Karine Emsellem, Lucie Cadorel. Studying urban space from textual data: Toward a meth-odological protocol to extract geographic knowledge from real estate ads.. O. Gervasi; B. Murgante; S. Misra; A.M.A.C. Rocha; C. Garau. Computational Science and Its Applications – ICCSA 2022 Workshops. Proceedings Part II, 13378, Springer, pp.520-537, 2022, Lecture Notes in Computer Science, 978-3-031-10561-6. ⟨10.1007/978-3-031-10562-3_37⟩. ⟨hal-03925785⟩
27 Consultations
49 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More