Skip to content
Data

Counts, indicators, comparisons, and change over time

Use public data carefully. A count, rate, service total, spending figure, or housing outcome means little without its definition, geography, time period, and method.

The central data question is not simply “what is the number?” It is whether two numbers describe the same population, geography, period, and method well enough to support comparison.

Point-in-Time counts, HMIS service records, housing inventories, budgets, contracts, and program outcomes answer different questions. Bay Evidence keeps those distinctions visible.

Local 2026 Point-in-Time totals, where a county released one

Bay Area counties · January 2026 counts; Alameda primary release, other rows July synthesis · People counted · local estimate, not a census

Verified through Alameda: 2026-10-19; other rows: 2026-07-24

Six-county synthesis: about 23,000 · prior synthesis about 25,700 · not comparable bar-for-bar with the county rows

Open regional source synthesis ↗

Alameda primary record · October 19 correction

Source: Alameda County 2026 Point-in-Time Count report (full report, October 2026); regional synthesis for other counties. Separate from the standardized HUD CoC series.
Methods differ by county. San Francisco changed enumeration timing and is not shown as a standalone bar because the current local source note does not carry one. The six-county regional figure should not be added to the bars below.

Indicators

Definitions, units, source systems, update frequency, and limits for the measures used across Bay Evidence.

Open →

Comparisons

Cross-county and cross-jurisdiction comparisons with clear warnings when measures are not directly comparable.

Open →

Source Explorer

See how verified sources are distributed by year, publisher, geography, and broad source type.

Open →

Downloads

Export the verified public source list as CSV or JSON for research and independent analysis.

Open →

Data Quality

Understand revisions, exclusions, missing data, methodology changes, and limits on comparison.

Open →

Geographic visualization will return only when Bay Evidence has an actual map product worth publishing. The former placeholder route now redirects here rather than presenting an unfinished room.