Geographic information from georeferenced social media data

Abstract

In the last few years, we are witnessing an emerging class of communication and information platforms some call social awareness streams (SAS) [15]. Available from social media services such as Facebook, Twitter, FourSquare, Flickr, and others, these hugely popular platforms allow participants to post streams of lightweight content items, from short status messages to links, pictures, and videos, in a highly connected social environment. Many of these items are associated with location coordinates in the form of latitude and longitude, or with a business or venue that is in turn associated with a precise location. The number of "geotagged" items is likely to grow with the number of people using geo-enabled devices to access and produce SAS data. The vast amounts of SAS data offer unique opportunities for understanding local communities and people's attitudes, attention, and interest in them. Robust methods for learning from SAS data about geographies and local communities, using methods from Artificial Intelligence, Information Retrieval and Natural Language Processing, can greatly improve the state of geographic information retrieval. Such contributions, from better modelling of geographic areas, to improved knowledge about these areas and how they are used by individuals and communities, have begun to surface in the last few years, and are summarized in this article. The structure of this work borrows from Lynch [13], who referred to five elements that make up an individual's perception of city: districts, landmarks, paths, nodes, and edges. This article borrows from Lynch in the context of social media and SAS, proposing the four elements that make up the geographic information that can be derived from social media about a city or area: districts, landmarks, paths, and activities. This article presents a simple model for geographic SAS data, and then considers the four social media elements, or main types of applications of SAS data to geographic information systems. These applications include boundary definition and detection (district); computation of attractions (landmarks); derivation and recommendation of paths; and evaluation of activities, interests and temporal trends.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 41 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.