Methodology

​Data Sources

​

State Finance

U.S. Census Bureau — State Government Finance Time Series (GS00SG01)
File: GOVSTIMESERIES.GS00SG01-Data.csv (public release).
Coverage: 2012–2023; Geography: 50 U.S. states + national totals.
Note: District of Columbia is excluded from the 2023 finance release because of inconsistent function coverage.

​

Homelessness

HUD — Point-in-Time (PIT) counts and related homelessness data.
Source tables are Continuum of Care (CoC)–level; aggregated to state totals using HUD’s CoC-to-state mapping for the given year.
HUD metrics appear on “population/homeless” views and are not used in computing budget shares.

 

Social & Living Indicators

 

ACS Data Sources (1-Year Estimates)

Income, Housing, and Costs

• B19013_001E — Median household income

• B25064_001E — Median gross rent

• B25077_001E — Median home value

• S2503_C03_001E — Rent burden: % of renter households paying ≥35% of income

Poverty & Employment

• DP03_0128PE — Poverty rate (% of people below poverty level, past 12 months)

• DP03_0009PE — Unemployment rate (civilian labor force, % unemployed)

Education

• DP02_0068PE — Educational attainment: % of adults (25+) with bachelor’s degree or higher

• DP02_0059PE — School enrollment: % of population 3+ years enrolled in school

Population

          DP05_0001E — Total population

​

Crime

Annual state series (current product):
FBI Crime Data Explorer (CDE) — Estimated Crimes (State-Year).
File used: estimated_crimes_1979_2024.csv (public distribution).
Coverage: 1979–2024; Geography: 50 states + District of Columbia + U.S. national line.
Offenses included: violent crime total, property crime total, and components (homicide, rape (legacy & revised), robbery, aggravated assault, burglary, larceny, motor-vehicle theft). Arson is not present in this file.

​

 

Processing (State Finance Shares)

 Select direct-general function rows & denominator.
From GS00SG01, keep:

• Function dollars: labels beginning General Expenditure, by Function:

• Denominator: Total Expenditure – General Expenditure – Direct General Expenditure.

 Harmonize to Census functional categories

• Public Welfare

• Education

• Health & Hospitals

• Highways

• Corrections

• Police Protection

• Financial Administration

• Natural Resources

• Parks & Recreation

• Interest on Debt

• Other

 Compute per state–year (2012–2023)

• Amounts ($): published function dollars.

• Shares (%): function_dollars ÷ Direct General Expenditure × 100.

• Shares rounded to one decimal place in outputs.

 National (United States) line
Sum state dollars by function and sum the Direct General denominator, then compute national shares = national_function ÷ national_total × 100 (a     composition, not an average of state percentages).   

Note on Shares: Functional budget shares are rounded to one decimal place. As a result, totals may not equal exactly 100%. This rounding convention follows common statistical reporting standards.

 

Processing (Crime)

 

Annual state series

  Input columns (from CDE Estimated Crimes): year, state_abbr, state_name, population, violent_crime, homicide, rape_legacy, rape_revised, robbery,                aggravated_assault, property_crime, burglary, larceny, motor_vehicle_theft.

  Transformations

• Totals:

  • violent_total = published violent_crime. Where needed for consistency checks, we treat rape = revised if available, else legacy.

  • property_total = published property_crime (burglary + larceny + motor-vehicle theft).

• Rates (per 100,000): count ÷ population × 100000 for violent and property totals (and components where shown).

• National line: population-weighted composition: for each year we sum state counts and sum population, then compute national rates and totals.

• District of Columbia: included in crime products as its own state‐equivalent; national uses all 50 states + DC. (Finance national line excludes DC per Census finance coverage note.)

Outputs

• CSV (state-year): totals + rates.

• Charts: Violent vs. property rate over time; property subcomponents (burglary, larceny, vehicle theft) over time.

Important notes & limitations

• SRS → NIBRS transition (circa 2021): agency participation and imputation vary; some states show measurement breaks or increased volatility.

• Rape definitions: legacy and revised series coexist across years; totals prefer revised where available, else legacy.

• Arson: not present in the Estimated Crimes state-year file used here.

• Revisions: FBI periodically updates estimates; later downloads may differ slightly from earlier releases.

 

Processing (Homelessness)

  HUD homelessness: Aggregate CoC-level PIT to state totals; join by year to the state table for display. HUD values do not affect finance shares.

  ACS indicators: Pulled via API and joined by state FIPS; displayed alongside finance shares and Living Data.

 

Quality Assurance

• Finance shares per state–year sum to ≈100% (allowing for rounding).

• Crime totals: internal checks that property ≈ burglary + larceny + motor-vehicle theft and that violent ≈ homicide + rape (legacy/revised) + robbery + aggravated assault; flag and annotate exceptions coming from source definitions.

• Large YoY/MoM deltas flagged and spot-checked against source series (finance & crime).

• Dollars are nominal (no inflation adjustment).

• Shares describe composition, not service levels or policy outcomes.

• Territories and DC: excluded from finance calculations; DC is included in crime products as a state-equivalent and can be filtered in UI/exports.

​

City Social Data (2010–2024)

Sources
Data for U.S. cities are derived from the U.S. Census Bureau’s American Community Survey (ACS) 1-Year Estimates, the Bureau of Labor Statistics (BLS), and HUD’s housing indicators, harmonized for consistency across years and geographies.
Coverage includes all U.S. cities and census-designated places with population ≥ 65,000, spanning 2010–2024.

I

ndicators
The following indicators are extracted annually for each city:

• Education

  • High School Enrollment (%) — school enrollment share of population ages 14–18 (DP02_0059PE and derived ACS cross-tab).

  • High School Graduation (%) — share of adults (25+) with a high school diploma or higher (DP02_0058PE).

  • College Enrollment (%) — share of population 18–24 enrolled in college or graduate school (DP02_0068PE subset).

  • College Graduation (%) — share of adults (25+) with a bachelor’s degree or higher (DP02_0068PE).

• Poverty & Employment

  • Poverty Rate (%) — DP03_0128PE (population below poverty line, past 12 months).

  • Unemployment Rate (%) — DP03_0009PE (civilian labor force, % unemployed).

• Housing

  • Rent Burden (%) — S2503_C03_001E (renter households paying ≥ 35% of income on rent).

 

Processing

• Each ACS variable is retrieved via the Census API, rounded to one decimal place, and stored by place FIPS, city name, state abbreviation, and year.

• Where ACS tables report estimates as percentages, values are preserved as published.

• City-level series are aligned by year and filtered to maintain consistent population coverage across time.

• Minor gaps from sampling error or suppressed estimates are left blank rather than interpolated.

 

Outputs

• Annual city-level CSV/XLSX datasets (2010–2024).

• Charts showing trends in education, poverty, employment, and rent burden per city.

• Optional national and state averages for comparative visualization.

 

Limitations

• Some smaller cities intermittently fall below ACS 1-year publication thresholds; those years appear as missing.

• Estimates represent survey-derived percentages, not administrative counts.

• Rent burden and enrollment variables reflect Census Bureau methodology and may differ from local education or housing reports.

​

Attribution & Licensing

• Finance, HUD, ACS: per their public terms and documentation.

• Crime (annual): FBI Crime Data Explorer “Estimated Crimes” (public distribution).

• Include attribution and note any transforms when redistributing data packs.