Already housed or sheltered: what remains public

The preserved 2023 dashboard

1,021

“Already housed/sheltered” appears as its own category in the graph legend and the table’s total row.

This combines housing and shelter. The screenshot does not establish that these were unique people, nor does it identify their housing addresses.

It is a historical record of what the city publicly displayed. The cards and referral table are reproduced as shown; no percentage is calculated from them because their counting relationships have not been reconciled.

The supported finding: the breakdown shown here is absent from the current public Engagements view. The date and reason for that change are unknown. This does not establish that internal collection stopped.

What the current page displays

On October 21, 2026, we opened the Healthy Streets data and information page and selected Engagements in its Street Response Dashboard.

The Engagements graph has two series: total engagements and total referrals. It does not show the earlier “Already housed/sheltered” series. The sections have different reporting periods: a September 2026 encampment figure does not make the engagement figures current through 2026.

This describes the view observed on the review date. It is not proof that no equivalent graph exists elsewhere, or that the category was discontinued citywide.

Some of the detail survives in 311 notes

We searched public closure notes for explicit descriptions of existing housing or shelter and reviewed the matches. This is a conservative keyword-based minimum, not a complete census of outdoor contacts.

By request year: 14 in 2024, 9 in 2025 and 30 in 2026 through the available October 21 snapshot. Recent notes may be more explicit because of documentation and operational changes. These counts alone do not demonstrate an increase in the underlying population.

Examples that explain the distinction

• 711 Post: case 101004267403 describes dispatch finding an active assignment and facility staff confirming placement after transport. This is stronger evidence of an existing placement than an unverified self-report.
• 685 Ellis: case 101004509242 says three people remained in the area, all sheltered: one at a hostel and two at 685 Ellis Street. This is a staff statement in the closure note.
• Housing conditions: case 101002594244 says one person reported having housing but sometimes sleeping outside because of mold. That is a reported explanation, not an independently established inspection finding.

A 311 encounter address identifies where the response occurred. It must not automatically be treated as the person’s residence.

The 53 reviewed 311 cases

Exact public closure notes are reproduced below. Case links open the corresponding DataSF record. Search by location, note, case number or year.

53 of 53 cases

Method and limits

Window: requests dated January 1, 2024 through October 21, 2026, as available in the public snapshot. Request date is not necessarily the outreach date. Notes can be updated after a request is submitted.

Search coverage: the broad housing/shelter/residence query returned 50,000 rows, reaching November 2014, so it covered the selected 2024–26 request-date window for those keywords. Additional checks returned 434 all-history “housed” matches and 6,518 recent housing/apartment/lives-at/tenant matches. Many were irrelevant maintenance reports or false positives such as “unhoused.” This is not an all-history count.

Retained: notes explicitly describing existing housing, existing shelter or navigation placement, or a mixed existing SRO/shelter situation. Excluded: newly arranged or future placements, housing queues, generic policy statements, ambiguous timing and reports without a qualifying encounter. Two preliminary entries were removed because “was housed” or “housed in shelter” did not establish whether placement preceded the response.

Duplicates: case IDs were deduplicated, people were not. Six August 2024 cases have the same “Already sheltered” note. Some October 2026 group notes are nearly identical. They remain separate case records, not separate verified encounters.

No denominator: the 53 cases cannot be used to calculate the percentage of all outreach contacts already housed. The 2023 dashboard total and this extraction cover different periods, systems and counting units and must not be added or directly compared as a trend.

The separate Baker/Fell aggregate response, 17870623, describes 100 encounters across nine operations, including six already sheltered or housed. It is excluded from the 53-case count and is not treated as six unique people.

What Board records establish—and what they do not

The FY2023–24 Coordinated Street Response annual report, included in a Board packet dated October 22, 2024, describes an internal HSOC outcomes dashboard aggregating results by shift, including people already housed, people sheltered and people declining placement or information. It also defines the already housed/sheltered category using self-reports, recent referral knowledge or ONE System lookups. This documents internal tracking during that reporting period; inclusion in a Board packet is not proof that a supervisor discussed this particular category orally.

The FY2026–30 ICT plan, PDF page 30, describes ASTRID as a combined street-team database and its Outreach Worker Tool as completed, while a public-facing dashboard was still listed as to be determined at the time of that plan. This supplies context for data-system changes, not a documented explanation for the missing series.

No explicit HSH Oversight or Board statement announcing removal of this category was located in the searches for this page. To establish why and when it changed would require a dated dashboard revision record, a staff explanation in a meeting/report, or confirmation from the responsible department. To establish whether collection continues would require a current field definition, report or redacted export. The 311 notes establish that some current responses document existing placements; they do not establish that a complete citywide counter is maintained.

Sources and definitions

1. DEM Healthy Streets data and information — current dashboard page; interactive Engagements view inspected October 21, 2026. Historical image supplied by Bluoz.
2. DataSF 311 Cases and public JSON endpoint — per-case links are supplied in the table.
3. Official 311 dataset documentation — explains closure notes, privacy limits and multiple cases for one incident.
4. FY2023–24 Coordinated Street Response annual report, reproduced in the Board packet — its HSOC definition includes self-reports, staff knowledge of recent referrals and ONE System lookups showing shelter or PSH. Those are different evidentiary bases.
5. Piastunovich declaration, January 10, 2023 (82-21) and exhibits 82-22 through 82-25 — selected client logs with housing/status fields. These cover late 2022 and early 2023, not 2026. Their entries are not added to the 311 count.