Abstract:

In addition to satellite-based and airborne-based earth observation sensors, the citizen-based sensors have been recently playing its role in timely data acquisition. The term Volunteered Geographic Information (VGI) started its infancy stage in early 2007 when citizens have broad access to social media networks such as Twitter. In 2009, Twitter released the location service that enables its users to publish tweets with latitude and longitude. The availability of this ubiquitous and location-acquisition technology allows tagging of location information and hence, a collection of location-based Twitter data. However, noisy data from users through this media service need to go through the processes particularly text analytics and data visualization to make the data understandable and representable for scientific usages and sense-making. This paper surveys the techniques underlying the analytics processes in Twitter data by studying and comparing the techniques used in selected research papers in between 2011 and 2016, before producing the taxonomies for the techniques. Classification of the techniques is conducted by creating different sub-sets of different categorization. This taxonomy is to provide a classification of usages and its potential usages, particularly for location-based data. Opportunities and challenges are presented to ensure a successful development of any location-based systems in the future.