SOG: A Synthetic Occupancy Generator to Support Entity Resolution Instruction and Research

This paper reports on a project to develop SOG (Synthetic Occupancy Generator), a system to create realistic, but synthetic residential occupancy (name and address) histories as input for Entity Resolution (ER) processes. ER processes are intended to link records referencing the same, or related, real-world entities. Most organizations use some type of ER process to recognize their customers or clients across different channels of contact such as name and address, telephone number, or email address. However, growing concerns over customer privacy and identity theft have made organizations reluctant to publicly release personally-identifiable customer information. The result is that it can difficult to obtain actual occupancy information to use for student exercises or to experiment with entity resolution methods and techniques. SOG was created to address this problem by providing a tool capable of automatically generating a large number of realistic, but synthetic occupancy histories. SOG control parameters allow the user to customize certain features of the simulated occupancy histories. The project reported here is the first phase of a larger project. The second phase is to develop tools that will systematically disrupt the SOG output to create ER test files that have varying degrees of data quality and file formats.