The United States incarcerates more people than any other country on earth. That sentence has appeared in newspapers for decades, usually attached to a single national number: roughly two million people. The number is accurate. It is also almost meaningless.

Who gets incarcerated, and where, and for how long: those decisions are made almost entirely at the county level. Judges, prosecutors, and sheriffs operating under local rules with local budgets produce outcomes that vary so wildly across county lines that neighboring communities can have incarceration rates five times apart. The national figure averages all of that away.

For most of this history, the underlying data existed only in government archives, accessible to researchers with time, institutional access, and patience for large spreadsheets. The Vera Institute of Justice spent years compiling it into a single dataset. What follows is an attempt to make that data usable by anyone.

The map to the left draws on the Vera Institute of Justice Incarceration Trends Dataset, covering over 3,000 counties from 1970 through 2024. Each county's color reflects its jail rate: the number of people jailed per 100,000 residents. Darker means lower. Orange means higher.

What follows is what that data reveals.

The map is now animating through American history. Watch the orange spread from 1970 forward.

In 1970, most counties sit in the dark: low jail populations, sparse data, the country incarcerating roughly 160 people per 100,000. Then, over four decades, nearly every policy lever gets pulled in the same direction at once.

The 1973 Rockefeller Drug Laws established mandatory minimums for drug offenses that dozens of states would copy. Reagan's 1986 Anti-Drug Abuse Act created the 100-to-1 sentencing disparity between crack and powder cocaine. The 1994 Violent Crime Control Act funded 100,000 new police officers and $9.7 billion in new prison construction.

Since 2008, rates have declined. But the decline is uneven: some counties reformed decisively; others barely moved. The map makes that asymmetry visible in a way that national averages cannot.

The national boom did not land equally. In 1970, Black and white Americans were jailed at rates that, while already unequal, moved roughly in parallel. Then, through the 1980s, they diverged sharply and never recovered.

Holmes County, Mississippi, in the heart of the Mississippi Delta, represents the extreme end of this pattern. It is majority-Black, majority rural, and majority poor.

Nationally in 2019, Black Americans were jailed at roughly 3–4× the white rate. In counties like Holmes, that gap is far wider. It is the product of decades of specific local decisions: how bail is set, which offenses are prosecuted, how public defenders are funded.

The map has zoomed to Maryland. Statewide, its jail rate sits near the national median, which tells you almost nothing. Hover over any county to see why.

The variation between Maryland's highest and lowest county jail rates in a single year is larger than the difference between the United States and most European countries. These counties operate under the same state laws. The disparities come from local decisions about bail, prosecution, and diversion programs.

Dorchester County sits on Maryland's Eastern Shore, a rural, largely agricultural county of about 31,000 people. The median household income is roughly $45,000. It does not look like a place you'd expect to have one of the state's highest jail rates.

That rate is not explained by crime statistics alone. Dorchester has fewer public defenders per capita than Baltimore, fewer diversion programs, and a jail that provides some of the county's most stable public-sector employment. Once a correctional facility becomes a major employer, the incentive structure changes: elected officials become reluctant to shrink it.

Not everyone in a county jail has been convicted of anything. A growing majority are people awaiting trial who could not afford bail. The expansion of pretrial detention is one of the least-discussed drivers of the incarceration boom.

Nationally, the pretrial share of the jail population rose from roughly 40% in 1970 to over 65% by 2019. In some rural counties, it exceeds 80%.

A defendant who can post $500 bail goes home. One who cannot stays jailed for weeks or months, often losing their job, housing, and custody of their children before any verdict is reached. Bail functions as a penalty for poverty.

When people talk about mass incarceration, they typically mean men. The crisis is real and the numbers are large. But the fastest-growing segment of the incarcerated population since 1970 is women.

The female jail rate has grown roughly nine times faster than the male rate since 1970. The growth is concentrated in rural counties, particularly in Appalachia, where the opioid epidemic drove a wave of drug-related arrests that hit women disproportionately. Leslie County, Kentucky had the highest female jail rate in the Vera dataset in 2019.

This is not a phenomenon that shows up in national headlines. It lives in the county-level data, concentrated in places experiencing simultaneous economic collapse, opioid addiction, and the absence of treatment alternatives to incarceration.

After 2008, the national jail rate began its first sustained decline in American history. Reform advocates pointed to the numbers as evidence that the era of mass incarceration was ending. The map tells a more complicated story.

While many urban counties drove meaningful reductions, some rural counties continued to climb, sometimes dramatically. Tensas Parish, Louisiana saw its jail rate rise by over 4,500% between 2008 and 2019. The explanation involves a private detention facility that opened to house ICE immigration detainees, inflating the county's population-adjusted rate to an extreme.

The Tensas Parish case points to a pattern that the aggregate numbers hide entirely: after 2008, immigration detention became a significant and growing driver of jail population in rural counties with contracts with ICE. Karnes County, Texas, another private detention hub, shows the same pattern. These are not traditional criminal justice stories. They are revenue models.

Concho County, Texas has fewer than 3,000 residents. At its peak, it had one of the highest jail rates in the United States, not because of elevated local crime, but because it operates a regional jail that houses inmates from surrounding counties under contract. The jail is, effectively, an industry.

This pattern of rural counties building and expanding jails as economic development repeated across Appalachia, the Great Plains, and the rural South from the late 1980s through the 2000s. Hundreds of communities voted to fund new jail construction on the theory that correctional facilities would anchor local economies the way factories once had.

Most produced a generation of inflated incarceration without the promised economic benefits, leaving communities with facilities too expensive to close and too politically entrenched to shrink. The orange counties scattered across the map's rural interior are, in many cases, places where that bet was made and lost.

None of the eight patterns above are anomalies. The national boom, the racial divergence, the within-state variation, the pretrial expansion, the female incarceration surge, the uneven post-2008 decline, the rural jail economy, and the immigration detention complex all repeat, in different forms, across every region of the country.

Before this dataset, understanding any of these patterns meant filing records requests, parsing Bureau of Justice Statistics spreadsheets, or hiring researchers. The underlying data has existed for decades, sitting in government archives as rows and columns. What it lacked was a way to see it. This visualization puts 54 years of county-level incarceration data in one place and makes it navigable: click any state to explore its counties, click any county to see its full history broken down by race, gender, custody type, and jail vs. prison rate.

That matters because the numbers at stake are not small. The dataset covers more than 3,000 counties, roughly 54 years, and tens of millions of individual incarceration events. Each one is a person. Each county line you cross on the map represents decisions made by real officials under real local conditions that shaped real lives. The national figure collapses all of that into a single number. The map gives it back.

Data source
Incarceration Trends Dataset, Vera Institute of Justice (2023 release). County-level and state-level jail and prison population data assembled from Bureau of Justice Statistics Annual Survey of Jails and Deaths in Custody reporting records, 1970–2024. Available at github.com/vera-institute/incarceration-trends.

Geography
US county and state boundaries from us-atlas (Mike Bostock / Observable), rendered with D3.js and TopoJSON.

Design inspiration
Scrollytelling structure informed by Harry Stevens' Washington Post corona simulator and the MBTA visualization by Barry & Card. Color palette and panel layout adapted from Observable's dark notebook theme.

Libraries
D3.js v7 (Mike Bostock et al.) · TopoJSON Client v3 (Mike Bostock)

Built for CMSC 471: Introduction to Information Visualization, Spring 2026.
University of Maryland, College Park.