SocScape - High resolution demographic grids

We developed and made available three types of datasets: population grids, racial grids and racial diversity classification grids. Our data are available for each county and for 363 metropolitan areas for 1990, 2000 and 2010 year as multi-year comparable datasets (called “myc”). For 2010 we also made available additional datasets, that used more detailed ancillary information, but cannot be used for change analysis.


  • Spatial resolution:  30 m
  • Projection:  Albers Conical Equal Area (EPSG: 5070)
  • Data units for population/race grids: Estimated persons per grid cell
  • Data units for racial diversity grids: one of 40 classes of racial diversity classification
  • Nodata value for population/race grids: -9999
  • Nodata value for racial diversity grids: 255
  • Data format: Geotiff
  • Date of production: January 2016


Name of each zip archive contain 2-letters state code and county name. For example, indicates Hamilton county in the state of Ohio.

Each archive contains 3 directories:
(1) population  - contains population grids for each year (4 grids)
(2) diversity - contains racial diversity classification grids for each year (4 grids)
(3) race  - contains separate grids for 7 race/ethnicity groups for each year (27 grids)

Notice: non-Hispanic Native Hawaiian and Other Pacific Islander (nhpi) grid is not available for 1990 year because non-Hispanic Pacific Islander (nhpi) are included in Asian group in the U.S Census 1990.

Example: oh_hamilton_pop1990myc.tif
Filenames consist of several elements: 2-letters state code (oh), county name (hamilton), name of map (pop1990myc), file extension (.tif) .  The information about available datasets and its name is here.

Zip archives contain set of available demographic grids for each county. Zip archives contain all grids regardless of county boundary changes, if county name was the same for all years. For counties which were added/deleted in particular year, archives contain only data for years for which a county existed under this name

Archives, that include only part of data are:

  • co_broomfield (only 2010)
  • fl_dade (only 1990); data for 2000 and 2010 are in fl_miamidade.
  • fl_miamidade (only 2000, 2010); data for 1990 are in fl_dade
  • mt_yellowstonenationalpark (only 1990)
  • va_cliftonforgecity (only 1990 and 2000)
  • va_southbostoncity (only 1990)

For substantial changes to counties and county equivalent entities see U.S Census Bureau website (


Demographic grids have been calculated from census blocks using dasymetric modeling technique.

Dasymetric modeling disaggregates aggregated population data (provided by US Census) into grid cells using ancillary data that correlates with population density but which has a higher resolution.


Fig. 1. How dasymetric modeling works?

High resolution grids provided by SocScape project are the results of dissagregating US census block level data using National Land Cover Dataset (NLCD).

For multiyear comparison datasets we used National Land Cover Datasets 2001 and 2011 reclassified into 3 classes (urban, vegetation, uninhabited) and NLCD 1992/2001 Retrofit Land Cover Change Product reclassified into 3 classes (urban, vegetation, uninhabited).

For 2010 grids ancillary data includes National Land Cover Datasets 2011 (NLCD 2011) (reclassified into 6 classes) and National Land Use Dataset 2010 (NLUD2010, Theobald 2014). 

Population grids are product of dasymetric modeling. We also used population-based dasymetric model to disaggregate race/ethnicity groups.  Seven race/ethnicity hi-res grids are available: non-Hispanic white (nhw), non-Hispanic black (nhb), non-Hispanic Asian (nhas), non-Hispanic American Indian (nham), non-Hispanic Native Hawaiian and Other Pacific Islander (nhpi), non-Hispanic other race (nhother), Hispanic (hispanic).

Racial diversity map shows spatial character of racial diversity across the U.S. We use our population/race hi-res grids to prepare three-dimensional classification of grid cells based on:

(1) racial diversity:  determined on the basis of standardized informational entropy (E):
(a) low diversity class (E<0.37 and dominant race more than 80%); (b) high diversity class (E >0.73 and a dominant race less than 50% population); (c) medium diversity otherwise.

(2) dominant race: non-Hispanic white, non-Hispanic black, non-Hispanic Asian, non-Hispanic  Native American and Hawaiian, non-Hispanic other race, Hispanic.

(3) population density: (a) low density (less than 3 people/km2), (b) medium density (3-30 people/km2), (c) high density (equal or greater than 30 people/km2). 

The result is a diversity–race–density classification of population cells into 40 categories (39 diversity–race–density and uninhabited class).  (Fig. 2.).

Fig. 2. Three dimensional classification of racial diversity.



  • A. Dmowska, T. F. Stepinski, P. Netzel (2017) Comprehensive framework for visualizing and analyzing spatio-temporal dynamics of racial diversity in the entire United States, PLoS ONE 12 (3): e0174993. paper (open access)
  • A. Dmowska and T. F. Stepinski (2017) A high resolution population grid for the conterminous United States: The 2010 edition. Computers, Environment and Urban Systems 61, pp. 13-23. paper, preprint
  • A.Dmowska and T.F. Stepinski, (2017) Mapping Changes of Racial Composition in the United States: 1990–2010. Annual Meeting of Population Association of America, Chicago, IL, April 26-29 2017 paper

Published on  December 11th, 2019


GIScience 2020 Workshop - workshop web page

Information theory provides a consistent framework for the analysis of spatial patterns - blog post.

sabre: or how to compare two maps? - blog post.

GeoPAT2: Entropy calculations for local landscapes - blog post.

Blog posts series introducing GeoPAT 2 - a software for pattern-based spatial and temporal analysis.

new version of a Web-based query-by-example tool for finding places with similar climate is available here.

now more detailed (30 m ) population and racial diversity grids for 1990, 2000, 2010 are available here.

Resources for mapping racial segregation and diversity are available here.

Resources for mapping racial segregation and diversity using historical census tract data (1910 -  2010) are available here.