Where structural harm becomes visible.
Oklahoma’s counties carry overlapping burdens — poverty and poor coverage and food hardship and distance from care, often in the same place at the same time. The Atlas measures those burdens together, county by county, and shows what each one is carrying relative to its neighbors.
Live federal feeds
Four federal APIs refresh this Atlas on top of the embedded release. Each is read in your browser directly from the publishing agency. Once the data is retrieved you can freeze it into the file below, after which the Atlas paints instantly from the embedded release and still refreshes from the live feeds on top — embedded values are marked EMBEDDED with their vintage, refreshed ones are marked LIVE. If a feed does not respond within twelve seconds, the embedded value stands and the report names which feeds answered. Nothing is ever estimated to fill a gap.
What the tools do
| Report Builder | One county on its own, or up to four side by side. Grant, board and policy formats. Choose indicators, weight the eight domains on the iRISE™ 0–100 scale, and print or save as PDF. |
|---|---|
| Solution Forge | Embedded in the Report Builder. Reads the county's measured burden and your domain weights together, then returns pathways at the individual, organizational and structural level, each naming the public evidence registry that documents it. Where burden is heaviest or a domain is weighted at 75 or above, it requires a structural pathway and marks individual-level work insufficient on its own. |
| County Map | All 77 counties colored by burden on the indicators you pick, from the same feeds the Report Builder uses. Click up to three counties to compare them and carry the selection into a report. |
| County Profiles | A single county in full: composite and domain scores, every indicator with its rank and vintage, the widest gaps against a comparison county, and the fields that require action. |
Report Builder — single-county analysis and county comparisons
Use the Report Builder for one county on its own or for comparisons between counties. Check one county for a single-region analysis; check more to compare them side by side. Every report leads with how it was completed, shows county comparison tables and domain gaps to scale, and marks each field that needs action from the organization, an individual, or the community.
Your report will appear here. Print or save it as a PDF from the buttons above once generated.
County Map
Oklahoma’s 77 counties colored by burden level on the indicators you choose, from the same live feeds the Report Builder uses. Hover for the county name and value; select up to three counties to compare them side by side and build a report from the selection.
County outlines not drawing?
The outlines come from the Census cartographic county file, served by public mirrors. If every mirror is unreachable from your network, download the counties file and load it here — TopoJSON or a GeoJSON FeatureCollection both work. The burden figures do not depend on it.
County Profiles
One county in full — composite and domain scores, every indicator with its value, rank and vintage, and the widest gaps against a comparison county. Built to drop into a community health needs assessment, a board packet or a funder narrative.
Choose a county and open its profile.
How to read the numbers
Where every number comes from
Each scored indicator is one of two kinds. An embedded value was obtained from the publisher’s release and frozen into the Atlas for all 77 counties; its vintage and freeze date are printed in every report and on the County Map, and it is marked EMBEDDED. A live value is retrieved in your browser from the publishing agency’s open API at the moment output is generated and is marked LIVE. Live values overwrite embedded ones as they arrive; if a feed does not respond within twelve seconds, the embedded value stands and the report says which feeds did and did not respond.
The four feeds are the Census Bureau’s Small Area Income and Poverty Estimates, the Census Bureau’s Small Area Health Insurance Estimates, the American Community Survey five-year estimates, and CDC PLACES. County boundaries are drawn from the Census Bureau’s cartographic county file.
County Health Rankings & Roadmaps and HDPulse publish annual files rather than APIs, so their indicators are embedded: the County Health Rankings 2025 Oklahoma release supplies most of the Atlas, and HDPulse supplies median family income, families below poverty and unemployment from the American Community Survey 2020–2024. Both carry their vintage in every report.
Where County Health Rankings suppresses a measure because a county’s counts are too small, that county is left unranked on that indicator. Ten measures are affected, infant mortality and homicide most heavily. A county is always scored on the indicators that exist for it, never penalised for the ones a publisher withheld.
No embedded value is typed by hand. Each is read straight from the publisher’s own release file and written in unchanged, and the freeze control on the home page rewrites the block from a live retrieval whenever you want it refreshed. No value is ever carried forward from a prior year, filled from a state average, or modeled by the Atlas.
Ranks and burden levels
Every indicator is ranked among the 77 counties, taking direction into account, so rank 1 is always the heaviest burden. Ranks fall into five bands: heaviest (top 20%), heavy, moderate, lighter, lightest. A county missing a value on an indicator is not ranked on it and is not counted as zero.
Domain scores and the composite
Indicator ranks aggregate into eight equity domains scored 0–100, where higher means stronger conditions and lower means heavier burden, and the eight domains aggregate into the composite iRISE™ index. A domain score is the mean percentile of the indicators that returned a value for that county in that domain, so a county with partial coverage is scored on what exists rather than penalized for what a publisher withheld. Domains with fewer than two returned indicators are reported but marked thin.
County comparisons
Every report compares the county with one other county rather than a statewide median: by default the nearest county by centroid distance from the Census boundary file, or any county you choose. The gap on each domain is shown to scale, and the widest gap is the report’s lead finding. The Oklahoma value is shown alongside for reference.
Suppression
When a publisher withholds a value because the count is too small or the estimate is unreliable, the Atlas shows it as suppressed. It is never shown as zero, and it is never estimated.
What the Atlas will not do
It will not invent testimony, budgets, partners or within-county subgroup rates. Every such field in a report is marked with who must act: Action required — Organization, Action required — Individual, or Action required — Community.
Reading Oklahoma in particular
Thirty-eight federally recognized tribal nations have jurisdiction in Oklahoma, and reservation boundaries do not follow county lines. County-level federal data cannot describe tribal citizens as a population or tribal nations as governments, and this Atlas does not claim to. Where a county sits inside a reservation boundary, the county figure describes everyone living there and should be read alongside data held by the nation itself, on the nation’s terms. The American Community Survey race and ethnicity counts are carried in every profile as context, never as scored burden.
Citations
Every source the Atlas draws on, generated from the indicator registry embedded in this file so the list cannot drift from the data.
Boundary file
U.S. Census Bureau, cartographic boundary file for counties, 1:10,000,000 scale, 2023 vintage, served as TopoJSON. Centroids used for the nearest-neighbor default are computed from this file.
Glossary
Terms used across the Report Builder, County Map and County Profiles, including the measures added in the County Health Rankings and HDPulse release.
How the Atlas works
| iRISE™ | The Intersectional Resilience Index for Social Equity. A measurement framework that scores a place on how burdens overlap rather than on one category at a time. |
|---|---|
| Composite index | The 0–100 aggregation of the eight domain scores for a county. Higher is stronger; rank 1 is the heaviest burden in the state. |
| Domain | One of eight areas of life the Atlas scores: economic security, housing stability, health care access, health outcomes, mental health and social connection, education and learning, food and material hardship, and environmental and community conditions. |
| Percentile | A county’s position among the 77 on one indicator, direction-adjusted so that 0 is the heaviest burden and 100 the lightest. |
| Burden band | The five-level grouping of ranks used on the map and in reports: heaviest, heavy, moderate, lighter, lightest. |
| Direction | Whether a high raw value means a heavier burden (a poverty rate) or a lighter one (a homeownership rate). Every indicator carries its direction so ranks are comparable. |
| κ (kappa) | The compounding factor for overlapping burden. κ = 1 + 0.12 × (domains weighted at 50 or above − 1). It expresses that burdens carried together are heavier than the same burdens carried separately. |
| Weight | What you tell the Atlas about how much a domain drives your problem, on a 0–100 scale. Weights shape the report; they do not change the underlying data. |
| Comparison county | The single county a report is measured against — the nearest by centroid distance unless you choose another. |
| Action required | A field the Atlas will not fill because it belongs to people, not to published data: testimony, budgets, partners, lived experience. Marked for the organization, an individual, or the community. |
| Thin domain | A domain for which fewer than two indicators returned a value for that county. Reported, but flagged so it is not read as firm. |
Where a number comes from
| Embedded value | A figure taken from a publisher’s annual release and stored in the Atlas for all 77 counties. Marked with its vintage, such as 2025. County Health Rankings and HDPulse are embedded because neither publishes an open API. |
|---|---|
| Live value | A figure retrieved in your browser from a federal API at the moment output is generated, marked LIVE. Live values replace embedded ones as they arrive. |
| Awaiting | An indicator with no value yet in this session. Excluded from domain scores and the composite, and reported as missing rather than estimated. |
| Suppressed | A value the publisher withheld because the count is too small or the estimate unreliable. The county is left unranked on that indicator — never shown as zero, never estimated. Infant mortality, homicide, drug overdose, child mortality and residential segregation are the most heavily suppressed measures in Oklahoma. |
| Context measure | An indicator carried for interpretation and never scored as burden: age structure, rural share, and the race and ethnicity counts. |
| Vintage | The year or year range a figure describes, which is not the year it was published. Printed beside every indicator. |
| Margin of error | The range around a survey estimate. Small-county estimates carry wide margins; treat narrow gaps between small counties as inconclusive. |
| Age-adjusted rate | A rate standardized to a common age distribution so counties with different age profiles can be compared. |
| Crude rate | A rate as observed, without age standardization. The Atlas uses crude rates for social-need measures and age-adjusted rates for health conditions. |
Measures that need a definition
| Years of potential life lost | Years of life lost before age 75, per 100,000 residents. A death at 25 contributes 50 years; a death at 70 contributes 5. It weights early death more heavily than a death rate does. |
|---|---|
| Premature age-adjusted mortality | Deaths before age 75 per 100,000, standardized for age. |
| Preventable hospital stays | Hospital stays for conditions that good outpatient care should have headed off, per 100,000 Medicare enrollees. A high rate signals primary care that is missing or out of reach. |
| Income inequality ratio | Household income at the 80th percentile divided by income at the 20th. A ratio of 5 means the better-off fifth earns five times what the worse-off fifth earns. |
| Gini index | A second inequality measure running 0 to 1, where 0 is perfect equality. Carried alongside the ratio because the two can disagree about which county is more unequal. |
| Gender pay gap | Women’s median earnings as a share of men’s for full-time year-round work. 0.75 means 75 cents on the dollar; higher is closer to parity. |
| Food environment index | A 0–10 score combining access to a grocery store with income-based ability to buy food. 10 is best. |
| Limited access to healthy foods | The share of residents who are both low-income and far from a grocery store — more than a mile in a city, ten miles in the country. |
| Severe housing problems | The share of households with at least one of: overcrowding, high housing costs, no kitchen, or no plumbing. |
| Severe housing cost burden | Households spending more than half their income on housing. |
| School funding adequacy | Spending per pupil above or below what research estimates that district needs to reach national average outcomes. A negative figure is a shortfall in dollars per pupil. |
| Grade-level equivalent | Average test performance expressed as a grade level, so counties can be compared without reference to a single state test. |
| Segregation index | A 0–100 measure of how unevenly two groups are distributed across a county’s schools or neighborhoods. 0 is complete integration, 100 complete separation. |
| Disconnected youth | People aged 16 to 19 who are neither in school nor working. |
| Social associations | Membership organizations per 10,000 residents — civic groups, religious organizations, sports clubs, labor organizations. A proxy for places people gather. |
| Frequent mental distress | Adults reporting 14 or more days of poor mental health in the past month. |
| Frequent physical distress | Adults reporting 14 or more days of poor physical health in the past month. |
| Provider rate | Clinicians per 100,000 residents. Higher is better. Shown as a rate rather than the ratio so that counties with no provider at all can still be ranked. |
| Traffic volume | Average annual vehicles per meter of major roadway near where people live — a proxy for exposure to noise, crashes and roadside pollution. |
| Adverse climate events | Three separate measures: cumulative days above 90°F, weeks in moderate or greater drought, and federal disaster declarations. |
| Census participation | The share of households that returned the census form on their own. Low participation means a county risks being undercounted in the formulas that distribute federal money. |
Reading Oklahoma
| Tribal jurisdiction | Thirty-eight federally recognized tribal nations hold jurisdiction in Oklahoma, and reservation boundaries do not follow county lines. County data cannot describe tribal nations as governments or tribal citizens as a population. Where a county sits inside a reservation boundary, read its figures alongside data held by the nation itself, on the nation’s terms. |
|---|---|
| Rural share | The share of residents living outside an urban area. Carried as context because many indicators — driving alone, provider rates, traffic volume — move with it for reasons that are not burden. |
About
Dr. Crystallee Crain
Dr. Crystallee Crain is the founder of iRISE™ Data and the creator of the iRISE™ algorithm — the Intersectional Resilience Index for Social Equity — a measurement framework built on the premise that communities should have access to the data used to describe their reality. She built the Michigan Prevention Data Atlas, the free, open-access platform that serves as the proof of concept for the national iRISE™ site, and this Oklahoma Atlas extends that same open access to all 77 Oklahoma counties.
Through her research and consulting practice, Prevention at the Intersections, founded in 2010, she has spent more than fifteen years working at the intersection of violence prevention, public health, participatory research, and community power building. Her work is grounded in community-based participatory research principles, equity, and deep systems change. This has included serving as Co-Principal Investigator on the Washington State LGBTQ+ Survey in partnership with Washington State University and the Washington State LGBTQ Commission.
She holds a Ph.D. in Transformative Studies from the California Institute of Integral Studies, and has more than eighteen years of university teaching experience. She is an award-winning author, poet, visual artist and publisher, and hosts the W3 Award-winning podcast Us with Dr. Crystallee Crain.
Book a consultation with Dr. Crain · irisedata@gmail.com
Support the Atlas
The Oklahoma Prevention Data Atlas is free and open access for all 77 counties, with no paywall, no login and no data resold. Donations cover the research, the live federal feed maintenance and the hosting that keep it that way.
Donations are processed securely by Zeffy, which passes one hundred percent of each gift through to iRISE™ Data. iRISE™ Data is fiscally sponsored by Fiscal Sponsorship Allies, EIN 85-0839183; gifts are tax-deductible.
Data governance
The Atlas reads only public, published, county-level federal data. Nothing you type into the Report Builder or Solution weights leaves your browser, and no county file is uploaded anywhere — the County Health Rankings CSV you load is read in the page and discarded when you close it. Community data governance for iRISE™ is overseen by the Community Data Governance Council.