Dynamic Land-Use Map Based on Twitter Data

Yuyun Yuyun,Fritz Akhmad Nuzir,Bart Julien Dewancker

doi:10.3390/su9122158

Abstract

Location-based social media allows people to communicate and share information on a popular landmark. With millions of data records generated, it provides new knowledge about a city. The identification of land use intends to uncover accurate positions for future urban development planning. The purpose of this research is to investigate the use of social networking check-in data as a source of information to characterize dynamic urban land use. The data from this study were obtained from the social media application i.e., Twitter. Three kinds of data that are prioritized in this research are check-ins (specific location), timestamps, and a user’s status text or post activities. In this study, we propose a grid-based aggregation method to divide the urban area. Two different approaches are compared—rank and clustering methods to group the place’s activities. Then we utilize time distribution frequency to attain the land-use function. In this case, Makassar City, Indonesia, has been selected as the case study. An analysis shows that the check-in activity and the method we proposed can be used to group the actual land-use types.

Highlights

Urban planning is a technical process in the formation, arrangement, and development of a city
The problem on urban land-use mapping is deciding upon the particular region for certain land use
Previous studies have been conducted to detect land use over time, such as the use of aerial photographs for mapping and quantifying the change in forest land-use patterns [1], remote sensing [2], geographic information systems techniques [3], and Landsat images via satellite, which provide an efficient means for land-use detection [4,5]

Summary

Introduction

Urban planning is a technical process in the formation, arrangement, and development of a city. Previous studies have been conducted to detect land use over time, such as the use of aerial photographs for mapping and quantifying the change in forest land-use patterns [1], remote sensing [2], geographic information systems techniques [3], and Landsat images via satellite, which provide an efficient means for land-use detection [4,5] These approaches have some weaknesses, such as the inability of numerous sensors to obtain data and information in cloudy areas. We find that some researchers use these devices for land-use identification—for instance, the demonstration of GPS data for discovering a region and sensing human activity [12], urban Wi-Fi characterization [13], land-use and landscape identification using cell-phone data [14,15,16] These models concentrate on a particular region in a specific area, the lack of information from this data [17] and difficult to identify the user's footprint. We combine the individual’s travel time spread on weekdays and weekends as the parameters to define the land-use

Related Work

Data Collection

Text Mining for Place-Name Identification