Project

Discipline

place descriptions; digital cartography; user generated content (UGC); map generalization; volunteered geographic information (VGI); geographic information systems (GIS)

Lead
Dieses Projekt wird Methoden entwickeln, um geographische Information für die Visualisierung auf Karten zu vereinfachen und zu abstrahieren, unter Nutzung von Information aus dem Web und aus Social Media, mit dem Ziel der Entwicklung von neuartigen ‘place-based maps’.
Lay summary
In der Ära des Web 2.0 sind benutzergenerierte Inhalte (BGI), also durch Benutzer erzeugte Information auf dem Web wie z.B. Flickr-Bilder oder Tweets, weit verbreitet. Ein wichtiges Element von BGI ist häufig deren geographischer Bezug. Daher wurden in der Forschung Techniken entwickelt, um Information über bestimmte Orte zu extrahieren, und solche Orte (engl. places) zu charakterisieren mittels Beschreibungen, die aus dem BGI stammen. So kann z.B. ein Spielplatz durch Aktivitäten charakterisiert werden (Fussbal spielen, grillen usw.) und durch Dienste (z.B. gibt es Feuerholz?). Einen weiteren Trend stellt die zunehmende Adaptionsfähigkeit von Karten dar, indem z.B. Information mit Bezug auf den aktuellen Standort einer/s BenutzerIn dargestellt wird. Dieser Trend hat zur Entwicklung von Methoden zur Rangierung und Empfehlung von Orten geführt, als Basis für die Adaption von Karten. Ebenso wurden Verfahren zur adaptiven Kartengeneralisierung entwickelt, welche die Adaption der visuellen Darstellung auf Karten erlauben. Das Hauptziel dieses Projekts ist es, diese zwei Strömungen zu verbinden und Methoden zu entwickeln, welche die Operationen der Kartengeneralisierung automatisch adaptieren lassen in Abhängigkeit der Eigenschaften bestimmter Orte, wie sie zuvor aus BGI extrahiert wurden. So soll eine neuartige Form von Karten möglich werden: ‘place-based maps’. Eine place-based map ist ein integriertes ‘graphisches Bild’ des Angebotscharakters (engl. affordance) und der Themen, welche für einen Ort (place) und eine Nutzergruppe interessant sind. Wir werden solche Karten realisieren, indem wir Ortsbeschreibungen, welche automatisch aus BGI extrahiert wurden, nutzen um die adaptive Generalisierung zu steuern, und so eine place-based map generalization zu erreichen.

Direct link to Lay Summary

Last update: 03.06.2014

Active participation

Number	Title	Start	Funding scheme

In the era of Web 2.0, user generated content (UGC) has become abundant. Since an essential element of such UGC is information with a geographic connotation - ranging from volunteered geodata on OpenStreetMap.org, to georeferenced and tagged pictures on Flickr.com, and to location check-in on social media sites such as Foursquare, Facebook and Twitter - research has developed a variety of techniques to extract information about places, and thus place descriptions, the descriptions of places provided by everyday people. For instance, the characteristics of certain places, the activities and services they afford, a qualification of these services, and other information that is relevant to specific user groups.As a further trend, maps have increasingly become adaptive. Adaptation to users and their context of use is an indispen-sable element of modern web and mobile information systems, in particular location-based services (LBS) and social media applications. This has fostered the development of methods for providing information about the importance of places (relevance ranking, recommender systems) as a basis for adaptation, as well as techniques for adaptive map generalization, allowing adaptation of content in the visual portrayal of maps.However, these two streams of research have largely evolved in separation from each other. Place descriptions have hardly been used in map visualizations; and conversely, adaptive map portrayal in LBS and similar systems has been restricted to the use of static information (e.g. existing attributes of points-of-interest, links to related web pages). Methods that harness the potential of UGC-extracted place descriptions in map visualization are missing and current generalization methods still typically lack methods which specifically account for place.The overall objective of the proposed project is thus to develop methods to make map generalization operations adap-tive to the peculiarities of particular places extracted from UGC, and thus achieve a new form of map: place-based maps. A place-based map can be understood as a summarizing, integrated ‘graphic image’ in the Bertinian sense (Ber-tin, 1983) of the affordances and themes in a particular locale that a community of users is interested in. We will realize such maps by linking place descriptions extracted from UGC and unstructured text to drive adaptive generalization, thus achieving place-based generalization in the portrayal of maps.In order to achieve this objective, the proposed project seeks to capitalize on the results of previous research by both applicants , in two subprojects, each of which represents a particular stream of research in the applicants’ groups.Subproject A will focus on the generation of place descriptions using UGC and unstructured text. These descriptions will address all three aspects of place as defined by Agnew (2012): location (placenames and associated regions), locale (the environment and associated affordances within such a region) and sense of place (the experiences offered by a place to a group and the shared conceptualization of a place important to that group). Place descriptions will take two forms - population of a schema designed directly to drive the generalization process with traditional geometric data, and richer, more textually based descriptions, requiring more novel methods of incorporation into place-based maps. Subproject B will take the triad of place description information extracted in Subproject A and develop processes that allow controlling the automated generalization operations such that place-based maps can be obtained. Work planned in this subproject will start with a design of the components of place-based map generalization (data types, data flows, generalization operators, and map designs). Procedures will be developed that allow integrating the geometries, seman-tics, and textual descriptions of place descriptions into a form that can be usefully displayed on a map. Specialized gen-eralization operators will be developed that can exploit place descriptions, with a focus on abstraction and schematiza-tion. And finally, coherent processes of place-based generalization will be built from these operators.