Research
Each entry here is a self-contained analysis with its data sources documented and its methodology open to scrutiny. Use the dashboards interactively or read the methodology page for a description of the underlying data quality framework.
Dashboard
Per-CoC and per-state homelessness rates over 18 years. Sheltered vs. unsheltered breakdowns, US choropleth map, sortable state and CoC tables, and rate-format toggles (per 10k, percent, 1 in N).
Open analysis →Analysis
Tests housing market conditions (rent burden, vacancy rate, renter share), eviction rates, federal funding, and labor market indicators against the CoC homelessness rate. Restricted to data-quality-vetted CoCs (n = 361). Housing-market variables dominate; unemployment is essentially uncorrelated.
Open analysis →Methodology
Per-CoC undercount risk score combining sheltered share, rural designation, year-over-year volatility, and federal funding per homeless person. Flags 29 CoCs whose reported homelessness rates likely understate reality.
Open analysis →Geographic decomposition
For each state, compares the combined homelessness rate of Major City CoCs against the rate of all other CoCs in the state. Surfaces which state-level rates are dominated by their major cities — and which are not.
Open analysis →Geographic decomposition
Per-state homelessness rates broken into Major City, Other Urban, Suburban, and Rural CoCs. National pattern: Major City CoCs run ~4× higher rates than rural CoCs, but state-level patterns vary widely.
Open analysis →