The Dimensions databases, both demographic and business, are intended for use in statistical models and for neighborhood description. Both Demographic Dimensions and Business Dimensions are modeling databases at the block group and higher levels of geography that are useful for creating statistical models, site signature reports, and general executive summary information. Unlike discrete neighborhood classification systems (e.g. Panorama), these databases provide continuous measurement scores across the dominant demographic and business components that differentiate neighborhoods.
Both databases are based on the well-known data reduction tool of Principal Components Analysis, in which the common patterns found within a large number of variables are reduced to a core set of discriminating factors.
Each Dimensions database is normally provided as a set of continuous variables which are minimally auto correlated and have a mean of zero and unit variance. Factors are useful in a broad spectrum of applications, including:
Demographic Dimensions, when used in conjunction with Panorama segmentation and other targeting tools, can yield significant improvements in direct marketing results. By fine-tuning a segmentation profile, subgroups of segments can be targeted effectively.
Dimensions are minimally correlated and are therefore performance and site location models. Statistical models developed using factors tend to be less prone to prediction error as a result of multi-colinearity. Note that while it is possible to utilize both Dimensions databases within a single model, the Demographic Dimensions and Business Dimensions variables are not statistically uncorrelated.
Factors can be used to effectively describe the dominant characteristics of neighborhoods for use in demographic reporting systems. Site “signatures” are easily defined and analyzed, since each of these factors is independent and reflect the dominant neighborhood differentiators. SnapSite includes a site signature model as one of its core analytical tools.
Demographic Dimensions was completely rebuilt in the 2013A release and now contains 26 dimensions, based fully upon 2010 Census and ACS data. Several
hundred input variables were used in the analysis, which are summarized below by type of variable and source year. Note that in many cases, both average (or
median) and distribution data were used (e.g. median age, % population age < 18, etc.).
Business Dimensions was built using the 2012
BusinessCounts database, with the source data from InfoBase. The entire dataset was utilized by computing employment and establishment density, average size and percent of employment by detailed NAICS codes, and land use classification.