Information Extraction and Visualization from Twitter Considering Spatial Structure

Abstract

Mobile social media represented by Twitter are expected to be a suitable source of data for analyzing human behaviour and statuses of locations. It seems that we can provide location-based information simply by spatially filtering archived data. However, there are several problems in terms of practical use. This research considers in particular problems that concern the relationship between data meaning and their spatial structures. With regard to Twitter, in general, the location from which a tweet is posted is attached to a geotagged tweet. For example, the location coordinates attached to the geotagged tweet “Heavy rain in Miura Peninsula” by NHK (Japan's public broadcaster) are not those of the Miura Peninsula, but of Shibuya in Tokyo (where NHK is located). Therefore, the tweet is not found by a spatial search around the Miura Peninsula or even Kanagawa Prefecture (where the Miura Peninsula is located). To resolve such problems, we propose a framework that distinguishes locations of interest and locations of activity. We propose a method for automatically classifying such locations and develop a data collection, classification, and visualization system based on this method.