Different datasets in Social Explorer support different geographic levels. The AI automatically selects a dataset only if the requested geography exists within that dataset. Because each data source has its own publication standards, sample-size requirements, and geographic constraints, eligibility varies significantly across ACS, Census, and non-ACS datasets.
General Rules
ACS 1-Year data is available only for large geographies. This includes states, large counties, large cities, and large congressional districts. Smaller towns, census tracts, block groups, and ZIP codes are never included in the ACS 1-Year because their populations are below the minimum threshold for statistical reliability.
ACS 5-Year data covers all standard census geographies. This includes small towns, census tracts, block groups, and ZIP Code Tabulation Areas (ZCTAs). ACS 5-Year is the most flexible dataset when detailed geographic analysis is required.
The Decennial Census provides full counts of the entire population and is available at all geographic levels, from the national level down to individual census blocks.
EASI Projections support a wide range of geographies, including ZIP codes, tracts, and block groups, because the data is modeled rather than sample-based.
Other datasets, such as crime, health, environment and elections vary in their geographic coverage. Some administrative datasets are available only at state or county level, while others extend down to tracts or block groups, depending on the source.
Coverage by Geography Level
Nation
Available in all major datasets, including ACS 1-Year, ACS 5-Year, Decennial Census, EASI and additional sources.
State
Available across all datasets. ACS 1-Year covers all states because they exceed the minimum population requirement.
County
Available in all datasets. ACS 1-Year is available only for large counties. All counties exist in the ACS 5-Year, Census, and EASI data.
Place or City
Available across datasets. ACS 1-Year is limited to large cities. ACS 5-Year includes all incorporated places regardless of size.
Congressional District
Available across datasets. ACS 1-Year applies only to large districts with sufficient population. All districts exist in the ACS 5-Year and Census data.
Census Tract
Not included in ACS 1-Year. Fully available in ACS 5-Year, Decennial Census and often in EASI projections. Some non-ACS datasets include tracts depending on the source.
Block Group
Not included in ACS 1-Year. Covered by ACS 5-Year and Decennial Census. Some environmental or vulnerability datasets also include block-group detail.
ZIP Code
Not available in ACS 1-Year. Available as ZCTAs in ACS 5-Year. ZCTAs (ZIP Code Tabulation Areas) in ACS datasets are not real USPS ZIP codes but statistical approximations created from address and census block boundaries. Because of this, ZIP-level results from ACS always refer to ZCTAs. True USPS ZIP codes are available only in modeled datasets such as EASI or in specific administrative sources. Some health, environmental or crime datasets also include ZIP-level data depending on the agency.
School District
Not available in ACS 1-Year. Available in ACS 5-Year and Decennial Census. Some datasets include school district boundaries depending on the year and source.
Why These Eligibility Rules Exist
Geographic eligibility is determined by several methodological factors:
Minimum sample size for reliability
Small geographies often do not have enough survey responses in a single year to generate reliable estimates. Pooling five years of responses (ACS 5-Year) increases statistical stability.
U.S. Census Bureau publication standards
The Census Bureau defines strict rules for how and when data can be released for specific geographies to ensure methodological quality and confidentiality.
Privacy protection and data suppression
When populations are small, certain variables may be suppressed or withheld to protect confidentiality.
Methodological differences across datasets
ACS 1-Year is restricted to large geographies (populations of 65,000 or more) to maintain statistical reliability since the data is based on one year of sampling.
ACS 5-Year pools five years of data, increasing sample size and enabling reliable estimates for small areas such as tracts and block groups.
The Decennial Census covers all population units, including the smallest geographical units, because it is a complete count rather than a sample.
Modeled datasets like EASI can estimate values for nearly any geography, including ZIP codes and custom areas, but these estimates rely on statistical modeling rather than direct enumeration.
Availability of geocoded administrative data
Crime, health, environmental, and other administrative datasets depend on how agencies collect, geocode, and publish their information. This may not align perfectly with census geography structures.
What This Means for Users
When analyzing small areas (tracts, block groups, small towns), the AI will typically use ACS 5-Year Estimates, Decennial Census data or EASI projections, since these sources support small geographies reliably.
For large areas (states, metro areas, large cities), the AI may use ACS 1-Year Estimates when the most recent data is needed.
ZIP code requests are usually answered using ACS 5-Year ZCTAs or EASI projections. ZCTAs approximate ZIP codes, while EASI provides modeled results for USPS ZIP codes.
Requests involving detailed administrative datasets depend on how those datasets are reported by source agencies, which often determines the finest geographic level available.
Understanding these eligibility rules helps users interpret why a specific dataset was chosen and ensures that results align with both the geographic and methodological constraints of demographic data analysis