FEMA отмечает жилищную помощь владельцам в более чем 15 000 записях Пуэрто-Рико

The dataset Housing Assistance Program Data - Owners v2 changes the angle: from public infrastructure (PA) to homeowners who reported damages under Individual Assistance. The OpenFEMA page explains that the set is generated by the IA reporting team to share Housing Assistance program data for homeowners, aggregated by state, county/municipality, and ZIP, based on claims from DR-1439 (November 2002). It contains aggregated data without PII. The endpoint is https://www.fema.gov/api/open/v2/HousingAssistanceOwners.

Documented core fields: disasterNumber, state, county, city, zipCode, validRegistrations, inspections and severity (totalInspected, bands of femaInspectedDamage...), approvedForFemaAssistance, totalApprovedIhpAmount, and breakdowns such as repairReplaceAmount, rentalAmount and otherNeedsAmount. The definition of a valid registration requires that the applicant be in a state/county declared for IA and have registered within the period designated by FEMA. Approved IHP amount ≠ cash on hand the same day; it is the approved amount reported in the aggregate.

Verified query: filter state eq 'PR' reported 15,095 rows added for owners in the metadata of September 15, 2026. Restricting to disasterNumber eq 4339 (María) the row count was 7,509. These rows are aggregations by geographic combination (municipality/city/ZIP), not necessarily “7,509 unique households” in the census sense; the reporter should read validRegistrations within each row. In a sample from Adjuntas under María, one row with a valid registration showed totalApprovedIhpAmount of $3,673.31 (with repair/replace, rental, and other needs broken down) and another $377.58. These are examples of granularity, not an invented island-wide total.

Why this dataset is gold for Borinquen: it allows comparing municipalities and ZIP codes in approvals and inspections without requesting personal microdata. It allows contrasting the discourse of 'FEMA didn't help anyone' with the aggregated evidence of valid records and approved amounts—and, conversely, detecting geographic pockets with a lot of inspection and little approval. It also forces distinguishing Housing Assistance from Other Needs Assistance within IHP, and not mixing these aggregates with Public Assistance PW or HUD's CDBG-DR.

Honest limitations: (1) it is aggregated; it is not useful for tracking an individual case; (2) it reflects the owners' program, not the renters' (there is a sister dataset Renters v2); (3) totalApprovedIhpAmount is not synonymous with bank disbursement verified in another source; (4) the absence of a row does not prove that no one in that ZIP applied for assistance — there may be aggregation rules or thresholds; (5) you must always anchor the correct disasterNumber (4339 María versus 4671 Fiona, etc.).

Recommended writing method: open Declarations Summaries to confirm if IA was designated in the area; open Housing Assistance Owners filtered by PR and DR; tabulate the top municipalities by validRegistrations and by totalApprovedIhpAmount, summing only with the explicit label of “sum of OpenFEMA aggregates as of date X”; compare with PA Details for the same DR and explain that they are different publics. If the report touches on long-term housing (reconstruction, buyouts, mitigation), add the HUD CDBG-DR/MIT layer, which does not reside within this FEMA IA dataset.

The dataset page specifies the update frequency as R/P1D (daily) and classifies the set under Individual Assistance, with keywords “housing assistance.” The institutional program URL points to the general page of FEMA’s individual disaster assistance. That tells the reporter two things: first, that a scrape from yesterday may differ from today’s; second, that the regulatory context resides in the IHP program, not in Public Assistance. Anyone who cites Housing Assistance Owners as if it were the inventory of schools or roads is looking at the wrong shelf of the OpenFEMA index.

The bands of inspected damage —no damage, between $1 and $10,000, between $10,001 and $20,000, between $20,001 and $30,000, and over $30,000— and the approval bands (1–10,000; 10,001–25,000; 25,001–maximum; more than totalMaxGrants) allow for a narrative of geographic inequality without microdata. One municipality may show many inspections in the low band and few max grants; another, the opposite. That does not prove fraud or neglect: it proves a distribution that must be explained with eligibility rules, insurance, ownership, and the IHP cap in effect for each disaster. Inventing an “Island average” without summing and labeling the field is bad practice; not summing and claiming that “there is no data” is also bad practice.

Relationship with CDBG-DR: when the IHP is exhausted or does not cover long-term structural reconstruction, HUD intervenes with Community Development Block Grant Disaster Recovery and Mitigation. Those funds do not appear in HousingAssistanceOwners. The transparency workaround, then, is dual: FEMA IA for the first tranche of owners; HUD for the second. Confusing both produces the headline that asks “where are FEMA’s billions?” when part of the ten-year housing money was never from FEMA. This note explains the first tranche and refers to the second.

In the Caribbean, the memory of María and Fiona is told in blue roofs and in registration lines. This dataset does not replace that memory, but it makes it auditable. Teaching the fields —validRegistrations, approvedForFemaAssistance, repairReplaceAmount— is teaching how to read the federal government in homeowner code. Cousin: 749 (AI supplement versus PA). Topic: economy. Primary source: OpenFEMA Housing Assistance Program Data Owners v2.