International Collaborative Research Project on Statistical Physics Modeling of Socio-Economic Phenomena

Research Summary

We are focusing on modelling socio-economic phenomena through the analysis of huge amounts of precise data. Such Big Data analyses allows us to construct mathematical models using statistical physics. Using this quantitative approach, we aim in the prediction and regulation of social phenomena. The main subjects of our research are financial markets, business firm trading networks, words in cyber-space and human flow patterns using GPS.


  1. Financial Markets

Drawing on foreign exchange market data, we are building models to predict human and AI trading behavior using original physics methods such as fluctuation-dissipation relations.

  1. Business Firms Trading Network

By analyzing over 1 million Japanese companies, we have constructed a “gravity model” to estimate the flow of capital between firms, therefore enhancing our understanding of business networks in Japan. The gravity model has been adapted to estimate the flow of money among regions in the Cabinet Office RESAS figures (Regional Economy Society Analyzing System). We are also developing a dynamic mathematical model of company growth.  

  1. Words in Cyber-Space

We are analyzing posts on social media such as blogs and Twitter to construct mathematical models that can explain phenomena such as rumor propagation, can quantify society trends, and accurately identify the occurrence and decline of booms.

  1. Human Flow Pattern using GPS

Through analysis of GPS data made available by smartphone users of certain apps, we are investigating the patterns of human movement through urban environments. We are developing a unique mathematical approach that exploits a potential force model to estimate global movement averages.