Dynamic Census Project

Develop a method to create a human mobility dataset by analyzing mobile phone data with various secondary data such as land use and transportation networks. For collecting training and validation data, field surveys were conducted. Results obtained through the method are called Dynamic Census that is the human trajectory data and gridded-map data, representing the spatiotemporal distribution of both mobile phone users and non-mobile phone users. The data are labeled with predicted demographic attributes. We believe it can be good supplement data for conventional population and housing census data with the information on population movement at the high granularity and high frequency. Considering that the data structure of mobile phone data does not vary a lot according to the region and country, developed method will be scalable in other parts of the world. We just finished a pilot project in Bangladesh and are preparing for Sri Lanka and Mozambique.