Red bubbles represent eviction statistics, which include the eviction filing rate, the eviction judgment rate, or the households threatened rate, depending on what statistic the user chooses. Bigger bubbles are bigger numbers. Smaller bubbles are smaller numbers.
The blue shading represents the concentration of the Census value you selected. For example, if you selected “population,” a darker blue means a higher population and a lighter blue means a lower population.
You can compare three locations at a time.
In our original map, you may see empty bubbles. In those cases, there is no available data for the selected year. Data for that location may exist for other years, however.
Selecting “auto” enables you to zoom through the geographic levels (states, counties, cities, census tracts, and block groups) automatically. You can also select your desired geographic level from the same dropdown menu.
You can search by state, county, or city. You can also search for any listed address in America.
You can embed the map in your website or blog using the provided embed code. Under the “share” section of the map & data page, click the link button to grab the code. You can then paste that code into your own site. The map will show the eviction and census data types you selected.
Reports, which are customized to display the same data and locations you’ve selected via the map and search, are designed as tools to support research and activism. You can download
Get More Data will take you to a page where you can search for and download large data sets for the entire United States.
Our primary data were collected manually from courts across the country. We also collected proprietary data from private vendors (see ABOUT THE DATA below). Because some areas are missing data over time in the courts and/or proprietary data, we used statistical models to “fill in the gaps” with modeled estimates. You can perform statistical analysis on the court data and the proprietary data but NOT on the modeled estimates. For more details, see our methodological report.
Because these data were collected manually from courts across the country, some areas are missing data over time. In other cases, statistical models indicated that there was too much variance in collection from one year to the next. In these cases, we filled in some of those gaps by “imputing” data based on a set of procedures listed in our Methodology Report (PDF). We use the imputed dataset in our maps, but we also wanted to give everyone the opportunity to download non-imputed data.
We support the last two major releases of Edge, Chrome, Firefox, and Safari.
Nearly 30,000 places—cities, towns, villages, and unincorporated areas—are included in our database.
Our rankings visualization shows only the top 100 areas based on your selections. Using the search bar, however, you can find the ranking for any place in America.
If you cannot find a place via the search bar, it may not be a place listed in the Census, or it may be a kind of place (like a state or county) that is not ranked in this tool.
We chose to categorize places in our Rankings (as Large Cities, Mid-sized Cities, or Small Cities & Rural Areas) to allow for more meaningful comparisons among places with similar populations.
Small Cities & Rural Areas are places with populations under 20,000. We rank 27,403 of these places. Examples: Clarksdale, MS; Fremont, OH
Mid-sized Cities are places with populations of between 20,000 and 100,000. We rank 1,858 of these places. Examples: Carson City, CA; Evanston, IL
Large Cities are places with populations of 100,000 and above. We rank 313 of these places. Examples: Los Angeles, NYC, Detroit
An eviction happens when a landlord expels people from property he or she owns. Evictions are landlord-initiated involuntary moves that happen to renters, whereas foreclosures are involuntary moves that happen to homeowners when a bank or other lending agency repossesses a home.
In our data, an eviction is defined as an eviction judgment issued to a renting home. If a renter at the same address has multiple eviction judgments, we only count the last one. For more information, please see our Methodology Report (PDF).
For more information about eviction and its effects, see our Why Eviction Matters page.
An eviction filing is the result of a landlord filing a case in court to have a tenant removed from a property.
Within our data, case information is recorded as separate actions with dates of each action. Our estimate of eviction filings is the total amount of cases each year with a first recorded action in that year.
Over the course of a year, a landlord may file multiple evictions against the same household. We count each eviction filing when calculating the eviction filing number and rate.
An eviction filing does not necessarily mean that the threatened household leaves its home. A common practice is what is known as “serial filing,” where a landlord issues a series of eviction filings against a single household.
We can use information in our case records to identify when a household receives multiple eviction filings in a given year. By counting the number of households who ever received a filing in a given year, we can calculate the households threatened rate. The eviction filing rate will always be higher than the households threatened rate.
An eviction rate is the number of evictions per 100 renter homes in an area. An eviction rate of 5% means that 5 of every 100 renter homes faced eviction in the selected area that year.
Information on the number of renter homes in an area comes from the U.S. Census and ESRI Business Analyst demographic estimates. We interpolate between 2000 and 2010 Census years for renting households to calculate yearly estimates for all geographies. For 2011-2016 and 2016-2018 estimates, we interpolate between 2010 Census counts and 2016 ESRI Business Analyst demographic estimates (see the Methodology Report).
We have accumulated 99.9 million records related to eviction. The Eviction Lab directly collected court records from 19 states and 8 additional counties. There were also some instances in which researchers outside the EL helped us obtain court records or aggregate filing counts in specific locations. We wish to acknowledge and thank: researchers at Flatland for Wyandotte and Johnson Counties in Kansas; Jonathan Pyle for Pennsylvania and Philadelphia County; Ben Schofield for Virginia; Tim Thomas for Washington, Kevin Fagan for San Francisco, and Kyle Nelson for California.
But many states either did not centralize their eviction data or were unwilling to release this information. Accordingly, the Eviction Lab then purchased more comprehensive datasets of public eviction records from two companies: LexisNexis Risk Solutions and American Information Research Services Inc.
We used estimates of renter-occupied households from the 2000 and 2010 U.S. Censuses and ESRI Business Analyst 2016. For further details on the renter-occupied household variable, please read our Methodology Report (PDF).
We used the following surveys for other demographic variables listed on the website and data download:
|Years Associated||Survey(s) Used|
|2005-2009||2005-2009 ACS 5-Year Estimates|
|2010||2010 Census for Population estimates; 2006-2010 ACS 5-Year for all other estimates|
|2011-2016||2011-2015 ACS 5-Year Estimates|
See our data dictionary for more details.
See the following table for figures from our National Estimates study. These are estimates, but we are reasonably confident that the true filing rate is within approximately 0.2 percentage points of the listed estimate, while the true households threatened rate is within approximately 0.6 percentage points of the listed estimate. Exact credible intervals can be downloaded here.
|Year||Renting Households||Estimated Filings||Estimated Filings Rates||Households Threatened (Estimate)||Households Threatened Rate (Estimate)|
If you are looking at our original map, you might notice that Arkansas, North Dakota, Alaska, and South Dakota do not have an eviction rate. These are states for which we do not have enough individual-level eviction records to establish the eviction rate with confidence. In these states, we can display reported statistics about eviction filings, however. For more details about how we handled data from different sources see our Methods page.
Users interested in the data for these states can look at our modeled data, which does provide estimates for those locations that do not share eviction data directly. These estimates have an uncertainty level that is presented as minimum and maximum values, which our models suggest are 95% certain to contain the true number.
On the map, we have indicated with the symbol Census tracts and block groups with filing and eviction rates in the top 1%. There are areas with high rates because housing insecurity is especially acute there. But there are other considerations as well.
Remember that filing and eviction rates are calculated by dividing the number of filings or evictions in an area by the number of renter homes. That means that very high rates could be the result of (1) a large number of evictions or filings or (2) a small number of renter homes in an area.
Take neighborhoods in Prince George’s County in Maryland. Because the way Maryland records eviction notices, it has a much higher filing rate than everywhere else, but this doesn’t mean it has more evictions. In most jurisdictions, the eviction process starts with an out-of- court notice delivered to a tenant, but in Maryland the process begins with an eviction filed in court. Many landlords file against their tenants every month, resulting in a very high case volume. Here, the number of filings is inflated because of unique court procedures, resulting in a high rate.
Now consider a residential neighborhood bordering a large industrial zone. The Census tract that includes the industrial zone also includes a small residential section with 40 renter homes. If there were 20 evictions from those homes over the course of a year, that would generate an eviction rate of 50%. Here, the eviction rate is high because the number of renter homes in the area is small.
There are even cases with filing and eviction rates over 100%. If there is a filing rate over 100% in a given area, it means there were more cases filed in a year than there were renting homes. Some eviction cases end without a tenant leaving the property. In some jurisdictions, tenants are taken to court repeatedly as a means of collecting past due rent.
But remember that high rates can be driven by the denominator, too: low renter homes relative to eviction cases. This might be because of the specific make-up of an area (e.g., rural community, residential area next to a major airport) or the issue of the Census undercounting renters in some areas. A residential neighborhood surrounding a college with 20 evictions but only 10 estimated renter homes would report an eviction rate of 200%. In this case, the high eviction rate is explained by the fact that the Census often doesn’t count seasonal renters (like college students) in its estimate of the number of renter homes in an area.
To get a sense of high rates, we recommend you look closely at the raw number of eviction cases and renter homes in your community.
On the original map, we have indicated with the symbol states with underestimated eviction counts. Of course, some places have low eviction rates because there are few renters living there or because there are very few evictions. But some states and the District of Columbia have underestimated eviction counts in our dataset for a variety of reasons.
Because coverage within our dataset is dependent on the area and year, it is possible for a single county to have fluctuations in data collection over time. As explained in our Methodology Report, we have flagged counties and embedded geographies (e.g., census tracts, block groups) with low eviction estimates. However, since we validated our data at the county level, counts of eviction filings at the level of census tracts and block groups that make up a low-estimate county might not in fact be undercounted. Our original data downloaded from the Get the Data page will contain these flags.
Here is a list of known issues with data collection at the state level, which have been flagged on our Orignal Data map:
In New York, records are often kept as “abstracted judgments,” which means they are only in the public record if the plaintiff/landlord pays to have them placed there. That plus the amount of town and village courts in the state makes collection difficult.
In California, many cases that end in eviction are sealed and therefore not accessible by the general public. In addition, it can be difficult to collect data from California as a whole, owing to restrictions on the number of records one can collect.
In most jurisdictions, the eviction process starts with an out-of-court notice delivered to a tenant, but in Maryland the process begins with an eviction filed in court. This means Maryland has a very high case volume. Because of that, it is difficult to collect data from Maryland, with the exception of Prince George’s county where we have consistency in case volume over time.
In New Jersey, while the number of eviction cases was collected reliably, information about the outcomes of those cases were not readily accessible. As a result, we believe our eviction rate is an underestimate.
In Kentucky, Louisiana, Tennessee, and Texas, there is good data coverage in the urban centers, but there are some more rural areas that are missing data, owing to collection difficulties. Additionally, due to a collection anomaly in Orleans Parish, Louisiana, possession-only judgments may be under-reported in this jurisdiction in 2016.
You can compare eviction rates from our map and raw data with court-reported statistics on eviction filings from 28 states and the District of Columbia (download court-reported statistics). These statistics provide annual counts of evictions filed in counties up to 2016. To keep things consistent, we haven’t included these statistics in our map of evictions, but you can use them to help validate the original data.
The court-reported statistics show that Hawaii, Vermont, Connecticut, Wyoming, and the District of Columbia have low counts in our original map and data, owing either to the remoteness of some areas or data collection difficulties.
We also suspect that the numbers in Arizona, Idaho, New Hampshire, and Washington may be too low, based on an estimator we created that factors in things like the number of renter homes and caseloads in other places.
We will continue our effort to provide the most comprehensive data on evictions in America and invite you to stay tuned for updates to our dataset by signing up for our email list.
It’s probably not because almost all eviction cases end in eviction but because of our inability to collect all eviction cases in your area. We attempted to collect every court record related to eviction—including filings that didn’t result in eviction—but some of our data sources prioritized cases that ended in a civil judgment. For example, in some areas, dismissals were not recorded, which deflated the total filings numbers. When this happens, filing and eviction rates are very similar.
While other units of geography are all connected (block groups, census tracts, counties, etc.), place boundaries are separate—requiring evictions to be aggregated and estimated differently. In many cases, the estimates of eviction are non-whole numbers (e.g., 0.4, 31.5, 100.6).
One feature of this strategy is that some low population areas could have less than one eviction estimated for that place. This should not be interpreted as half an eviction occurring in that place in a given year but rather as an estimate of the prevalence of eviction in this area. It is important to note that the rate displayed is based on the unrounded estimate as opposed to the rounded number displayed on the map.
Our estimates for “cities/places” are based on the Census definitions of places. For more details on how the estimation process for places differs, please see our Methodology Report (PDF). When downloading any of our “cities” files from the Get the Data page, you will receive the unrounded estimates of eviction/filings.
The Census Bureau considers a household rent burdened if it spends more than 35% of its annual income on rent. The measure we provide on the map tells you what percentage of renter households in a given area are rent burdened.
Since eviction records do not include demographic information like age or gender, we do not have direct access to these numbers. However, Eviction Lab’s researchers have published studies that show that Black and Latinx female renters face higher eviction filing rates than their male counterparts. Black and Latinx renters are also more likely to be serially filed against for eviction at the same address.
There are also local studies show that households with children are especially vulnerable to forced displacement. While we don’t have the full picture now, the Eviction Lab dataset enables users to create estimates on a scale that was previously impossible, and important questions like these can be addressed in future research.
We’ve collected over 99.9 million eviction records going back several years, but we still don’t have them all! To solve this problem, we have created estimates in counties where we lack data, which are created using a model. These estimates have an uncertainty level that is presented as minimum and maximum values, which our models suggest are 95% certain to contain the true number.
We will continue our effort to provide the most comprehensive data on evictions in America and invite you to stay tuned for updates to our dataset by signing up for our email list. If we don’t have eviction numbers for your community, you could help by contacting your mayor or county clerk, requesting this information. If you have eviction data we don’t, consider sharing it by writing us at email@example.com.
Data quality for eviction judgments is quite low, and so we are not able to offer national estimates of eviction judgments. Using our original data, we can observe eviction judgments in some states. The table below presents annual figures, but they should be interpreted with caution.
|Year||Eviction Judgements||Estimated Judgment Rate|
We have developed a Media Guide and Teacher’s Guide. We’ll be highlighting how journalists, teachers, community organizers, policymakers, and engaged citizens have used our data—both on our Facebook and Twitter accounts and our Updates pages. You can also download reading group guides for the book Evicted.
Sign up to receive news about events and new features from Eviction Lab.