I’m focusing on the data mining research and the social media analysis, especially topic extraction from millions of tweets related to the East Japan Great Earthquake. I’m also conducting global researches for developing the social media analysis platform in multi-language/cultural environment.

Time series topic extraction from social media

conventional / proposed method
Conventional Method and Proposed Method
conventional / proposed method
Analysis Example

Social media offers a wealth of insight into how significant events–such as the Great East Japan Earthquake, the Arab Spring, and the Boston Bombing–affect individuals. Analyzing social media lets us explore users’ reactions, thoughts and behavior in the reaction to events. The scale of available data, however, can be intimidating: during the Great East Japan Earthquake, over 8 million tweets per day were sent from Japan alone. Discovering when such an event occurs, and classifying tweets into those that are relevant to an event, remains a significant problem and an ongoing area of research. Many techniques have been proposed such as graph based methods, Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA), but no proposed method scales adequately to millions of tweets. Recently, however, we have developed a fast feature selection algorithm, namely CWC, that can apply to such large datasets. In this paper, we adopt it to discover and analyze events from a dataset of two hundred million tweets sent during the 21 days following the Great East Japan Earthquake. We form time series bipartite graphs of authors and words for each hour, and directly extract important features from each graph. Then we evaluate feature changes over time. Significant feature changes are then understood as events.

This research’s contribution is as follows:
* to propose the new event detection method from huge amounts of social media data
* to identify events from two hundred million of tweets after the Great East Japan Earthquake to analyze human reactions after the earthquake

Collaborative research groups

Language independent social media analysis platform

conventional / proposed method
Topics between Thai and Japan

After the global topics like a severe disaster (e.g. the East Japan Great Earthquake, Thailand floods etc.), people all over the world post their thoughts/opinions to social media and their behavior are sometimes influenced by the discussions on social media. For example, after the East Japan Great Earthquake occurred in Japan on 11th March 2011, many messages about the earthquake were posted to social media such as Twitter, Facebook, and so on, not only in Japan but also from all over the world. A lot of discussions were conducted concerning damages by the earthquake, affected people by the earthquake and so on, and the topics triggered many kinds of supportive movements. Various activities were activated by people who were affected by social media.

We believe that exploring topics related to the global topics on social media in different countries is useful to gain a rich insight into the global social contexts. We would like greatly to understand what are talked in the other countries. The goal of our research is to analyze different countries’ people reactions on global topics when disasters occurred such as the East Japan Great Earthquake.

Our method, first, crawls messages in social media and extracts keywords using the morphological technique. Next, we extract snapshot topics at each time stamp. Then, we investigate topic transitions over time by computing similarities of topics. Our method could show the time series topic transitions. Our method extracts topics based on keywords. Once keywords are extracted, the method can be languageindependent. The common language is supposed to be English. The keyword would be translated to an English word. We already tried a preliminary experiment to apply the method to other languages. As the first target, we selected Thai language. In Thailand, social media is quite popular. For example, after the East Japan Great Earthquake, a lot of people posted their messages about the quake on social media.

Collaborative research groups

Social media marketing in multi-cultural environment

We are conducting the social marketing research that targets Japanese contents such as Fashion, Comics and Foods in Asian countries using Facebook and so on. The aim of this research is to analyze the difference in reactions between Japan and other counties.

Collaborative research groups

Research towards Sustainable Societies

Under the Chiba Commerce University President Project , in order to realize a sustainable society, based on the concept of SDGs, we are conducting research on USR (University Social Responsibility, university social responsibility), 100% natural energy university, and distributed energy society.