Academic journal article International Journal of Electronic Commerce Studies

Social Sensor: An Analysis Tool for Social Media

Academic journal article International Journal of Electronic Commerce Studies

Social Sensor: An Analysis Tool for Social Media

Article excerpt


1.1 Research Motivation and Purpose

The rise of social media networks established a new style of network structure, and the policy of open data has led to data analysis of the global social media frenzy. The current rapid growth of a huge amount of social media data has led many people and resources invested in the task to collect, filter, store, process, and manage a huge amount of complex raw data quickly. In addition, the growth rate has been much greater than the speed that a human expert can analyze.

Ben Lorica1 pointed out that data scientists tend to use a variety of tools, often across different programming languages, to process big data. Workflows that involve many different tools require a considerable amount of context-switching, which affects productivity and impedes reproducibility. Similarly, the analysis workflow in a data science team usually adopts a sub-role, sub-field approach to the entire analytical work. In the team, some engineers are responsible for crawl data, and then transfer the data to processing data operations performed by other engineers. Output data is then explored and established by data analysts, which is then followed by data visualization staffto present analysis results to the domain expert for interpretation.

In this research, we studied the data analysis process of a research team at National Chengchi University who are analyzing social media. We used the case study "2012 Taiwan presidential election"2 to illustrate the problems in a typical analysis process. These challenges include the unmatched speed of data analysis with the speed of data collection, independent and fragmented analysis steps, labor intensive manual analysis, manual file exchanges, lack of data and case management tools, difficulty in maintaining domain expertise, high restart costs, long waiting times, etc.

As such, in this research we propose a new concept for social media analysis called "Social Sensor," which is an innovative design attempting to transform the concept of a physical sensor in the real world into the world of social media with three design features: manageability, modularity, and reusability. This tool is used to help users collect a variety of social media data for rapid analysis and visual presentation. By developing the concept of social sensors for a massive amount of social media analyses, we aim to provide an analytical tool for application in social science studies addressing policy making, marketing, and the design of an intelligence interface.

1.2 Concept and Design Goals

This research attempts to bring the concept of physical sensors to the virtual world of social media, shown in Figure 1. Physical sensors in a real-world environment are like social sensors in the social media environment, and they all produce a tremendous amount of environmental data. Physical sensors measure and produce physical or chemical data, such as light, heat, temperature, and humidity. The equivalent collected from social media by social sensors may include tweets, users, locations, etc. More specifically, examples of physical sensors include a luminosity sensor, temperature sensor, etc., whereas social sensors may include language sensors, text sensors and the like. Social sensors also function similarly to physical sensors because different sensors are designed to have their own sensing tasks and applications. A social sensor is implemented with various analytical methods and generates derived data for a specific analytical task.

In addition, some physical sensors may contain parameter settings that can be used to adapt to different applications, and the settings of these parameters usually come from prior knowledge and experience. Similarly, in the social sensor design, through parametric design, the same sensor can be customized to meet the needs of a specific application. Experience can then be passed on from one case to another in a different setting. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.