Who owns this building?

HPD classifies maintenance code violations into three severity classes. Class A is non-hazardous — peeling paint, missing window guards, stuff that's bad but not dangerous. Class B is hazardous — broken locks, defective plaster, things that can cause harm. Class C is the one that matters for our purposes: immediately hazardous. Lead paint. No heat in January. No hot water. Mice and roach infestations severe enough to constitute a public-health emergency. Fire egress problems. The kind of conditions that kill people.

The citywide breakdown since 2024 looks like this:

0 Class C violations citywide in roughly two years. That number sounds abstract until you concentrate it. The dozen worst buildings — each one with more than 20 Class C violations issued in the same period — look like this:

A building with 30+ Class C violations in two years is not having a bad month. It is being run in a way that is institutionally hostile to its tenants. The interesting question is: who's running it?

The corporate-veil problem

The named owner on each violation is, almost always, an LLC. The LLC was formed for the purpose of owning this one building. Its registered agent is a law firm. Its mailing address is a P.O. box. Its directors, on paper, are nominees. The actual beneficial owner — the person who collects the rents, who decides whether to make the repairs, who profits from running the building this way — is one corporate layer removed from anything you can find in the open data.

But the beneficial owner is findable. Tools like JustFix.nyc's Who Owns What, the Anti-Eviction Mapping Project, and the Right to Counsel Eviction Crisis Monitor do the join. They take this dataset (the violations themselves), the city's mailing-address registries (where the registered-agent and management-company addresses live), and the ACRIS deed records (which document the beneficial owner at the moment of property transfer), and they correlate. The same beneficial owner shows up behind dozens — sometimes hundreds — of LLCs. The same five attorneys keep filing for the same five property-management companies. The same handful of names keep appearing in the chain of title.

This is the corporate-veil-piercing pattern that tenant-advocacy work depends on. It is also the reason the city's open-data policy includes the violation data in the first place. The LLCs are obscuring; the violations are not. By making the violations queryable, joinable, and public, the city makes the obscuring strategy less viable.

What the data does not say

The dataset is conditions data, not causes data. A building with 30 Class C violations might have a negligent owner; it might have a building-systems problem that predates the current ownership; it might have tenants organized enough to file every violation they encounter. Volume of violations is jointly a function of physical conditions, owner responsiveness, and tenant civic capacity. Two buildings in similar physical states will accumulate very different violation counts depending on whether their tenants know to call HPD.

This is a feature, not a bug, when you remember what the dataset is for. The point isn't to identify the buildings with the worst conditions. The point is to identify the buildings the system is actually responding to. The Class C violation count is a measure of public action being taken against private negligence. When we list the worst buildings by Class C count, we're listing the buildings where tenants pushed back hard enough that the city pushed back too. Combined with the corporate-veil-piercing tools above, the data starts to name names.