Using neighborhood archetypes to understand unequal outcomes

Neighborhood data can make patterns visible. It can also make places look simpler than they are.

A poverty rate can tell us something. So can median income, racial composition, housing tenure, education, density, or employment. Each measure gives one piece of the picture. None of them, by itself, is the neighborhood.

Places are made from conditions that sit on top of one another. Housing, income, jobs, transit, segregation, services, environmental exposure, and institutional history all shape what people can reach and what they have to manage. They shape whether daily life is supported or strained.

That is why neighborhood archetypes are useful. Instead of asking one variable to stand in for a whole place, an archetype approach looks for recurring combinations of neighborhood conditions. It asks what kinds of places exist across a larger geography, and whether those different kinds of places are associated with different outcomes.

Disparities in prostate cancer survival according to neighborhood archetypes: A population-based study used this kind of approach to study survival differences across neighborhood types. The question was not simply whether poverty, race, housing, or another single neighborhood characteristic mattered. It was whether combinations of social and built environment conditions were related to survival.

Neighborhood archetypes and breast cancer survival in California asked a related question. Breast cancer survival is shaped by clinical care, stage at diagnosis, treatment, individual circumstances, and many other factors. But people do not enter those systems from nowhere. They live in places with different levels of access, stability, resources, and exposure to strain. Looking at neighborhood archetypes made it possible to study survival in relation to fuller neighborhood patterns.

Neighborhoods and breast cancer survival: The case for an archetype approach helped explain why this method matters. Neighborhoods are often studied through isolated variables because that is easier to model and easier to explain. But easier is not always more accurate. A single measure can point toward a problem while still missing the structure around it. Archetypes are one way to keep more of that structure in view.

Breast cancer in San Francisco: Disentangling disparities at the neighborhood level dealt with this problem in a more local way. Citywide numbers can hide sharp differences between neighborhoods. Even neighborhood-level patterns can be difficult to interpret when several conditions overlap in the same place. Disentangling disparities means looking closely enough to see where outcomes differ, who is affected, and what local context may help explain the pattern.

What ties this research together is a practical problem: how do we describe place without reducing it too far?

That is harder than it sounds. Neighborhoods are not just map units. They are not the same thing as census tracts, ZIP codes, or boundaries drawn for administrative convenience. Those boundaries are useful because analysis needs structure. But lived experience does not always follow them. People cross boundaries for work, school, medical care, food, transit, family, and services.

Official boundaries still matter, though. They shape which data exists, which programs apply, which agencies are responsible, and which decisions get made. A boundary can be imperfect and still have real consequences.

The work is not just to put a dataset on a map. The work is to ask whether the way a place is being measured is strong enough for the question being asked.

Weak descriptions of place can lead to weak conclusions. If a neighborhood is described only by income, the analysis may miss housing conditions, transportation barriers, racialized patterns of access, environmental burden, or the distribution of services. If it is described only by race or ethnicity, the analysis may miss economic variation, institutional investment, built form, or differences in access. If it is described only by a boundary, the analysis may miss how people actually move through the surrounding area.

An archetype approach does not solve all of that. No method does. But it lets the analysis ask a more honest question: which conditions tend to appear together, and how are those combinations related to outcomes?

That way of thinking carries beyond cancer research.

Small governments, schools, and nonprofits often work with place-based information. They may be looking at resident concerns, service requests, student needs, attendance patterns, program participation, infrastructure problems, survey results, or neighborhood-level outcomes. The same risk shows up there. It is easy to grab one available variable and let it carry more meaning than it should.

Communities rarely work that neatly.

A school attendance pattern may reflect bus routes, housing instability, family work schedules, health, safety, or trust. A cluster of resident complaints may reflect broken infrastructure, poor communication, uneven maintenance, or years of people feeling ignored. A nonprofit’s service area may look one way in a summary table and another way when addresses are mapped against distance, language, income, transit, or local institutions.

Analysis has to be careful here. The point is not to make a map look more sophisticated. It is not to add more variables just because more data sounds more serious. The point is to describe conditions clearly enough that the response has a better chance of fitting what is actually happening.

For Space for Us, that is part of the practical value of the work. Many organizations already have pieces of the information they need. They may have addresses, intake forms, service records, surveys, meeting notes, participation logs, public datasets, or staff knowledge that has never been written down cleanly. Those pieces do not automatically become understanding.

They have to be cleaned. They have to be organized. They have to be linked carefully, checked against what people know on the ground, mapped when location matters, and interpreted in a way that someone can actually use.

Neighborhood analysis is not only about where something happened. It is about what kind of place it happened in, what else is happening there, and what the pattern may be asking people to notice.

A place is more than a point on a map. Good analysis keeps enough of that complexity visible so the people making decisions can respond to what is actually there.

Related publications

Disparities in prostate cancer survival according to neighborhood archetypes: A population-based study
https://pubmed.ncbi.nlm.nih.gov/34303761/

Neighborhood archetypes and breast cancer survival in California
https://pubmed.ncbi.nlm.nih.gov/33577928/

Neighborhoods and breast cancer survival: The case for an archetype approach
https://doi.org/10.1007/978-3-030-18408-7_10

Breast cancer in San Francisco: Disentangling disparities at the neighborhood level
https://pubmed.ncbi.nlm.nih.gov/31548180/