Historical Population Informatics : Comparing Big Data of Family Trees and the U . S . 1880 Census for Migration Analysis

Throughout history, migration has been a key force in shaping and transforming humanity. Many cultural, political, economic, and demographic patterns in the modern world can be traced back to the process of human migration. However, historical migration data over a long time span (e.g., the past 300 years) at either local or global scales have been scarce and difficult to obtain. In the past human migration has been studied primarily through qualitative analyses of survey data and archival sources, which are often too generalized or too small to reveal the spatiotemporal complexity of human migration patterns. To address this challenge, we collect, process and analyze massive amounts of user-contributed family trees on the web, combined with fullcount historical U.S. census data, to obtain new datasets on human migration at a global scale over several centuries, with a focus on North America and Europe. This paper reports preliminary results of this on-going study, particularly focusing on the assessment of the representativeness of the family tree data, compared with the more complete and official census data of 1880, which is a fullcount data (having each individual) with names, locations, and parents’ locations.