The lack of data upon which to base private- and public-sector decisions is an important challenge in rural areas of the United States and other countries. Data suppression, done to protect the confidentiality of individual firms and people, compounds the more general paucity of rural economic data in the US. This article evaluates the use of linear programming for estimating suppressed values in the important US Census data series County Business Patterns (CBP). The full CBP dataset, enhanced with the estimates for suppressed values, was then used in shift-share analysis (SSA) at several levels of spatial aggregation. Finally, a publicly available website was developed to map the results of the SSA as well as the values of fifteen variables that may affect regional competitiveness. The overall result is a methodology for mining, analyzing, and visualizing rural economic data.
Read full abstract