Use the Data!

Download the Data and Documentation

Downloading Raw Data:

The data powering this dashboard is publicly available:

All of the Urban Institute’s Upward Mobility Framework Mobility Metrics are publicly accessible and can be downloaded on the Urban Institute’s Data Catalog.
Inclusive Recovery Indices data can be downloaded via the data download button in the Inclusive Recovery Index section of the Data Dashboard page.
Pre-2020 Inclusive Recovery Indices data also can be downloaded from the Urban Institute website.
Supplemental data sources are available at the following locations:

Downloading Data in Charts:

All charts and maps in the dashboard include a download button that lets you save the underlying data as a CSV file.

For downloading images of the charts, there are two options depending on the chart:

The camera icon in the toolbar: For most charts, when you hover over the chart or click on it, a toolbar appears in the chart’s top-right corner (see image below). Clicking the camera icon will save the chart as a PNG image. Note that if you interact with the chart, by zooming in, filtering, or otherwise adjusting the view, the saved image will reflect those adjustments.

Image download button below the chart: The maps and a handful of charts include a button that saves a static image of the chart as a PNG file.

About the Data

Data Sources:

Unless otherwise noted in its “Source and Notes” drop-down, all data in the Economic Development, Human Development, and Neighborhood Development sections of the dashboard come from the Mobility Metrics. The data from the Fresno Demographics tab comes from the American Community Survey’s 2008-2012, 2013-2017, and 2019-2023 5-year surveys. The data powering the Inclusive Recovery Indices tab comes from two sources: pre-2020 calculations were downloaded from the Measuring Inclusion in America’s Cities feature on Urban’s website (Christina Stacy, Brady Meixell, Ananya Hariharan, Erika Poethig, and Solomon Greene 2020). Urban Institute researchers performed the same calculations on updated 2020 data for this project.

For more information on the Mobility Metrics from the Upward Mobility Framework used in this dashboard, please see the following resources:

Data Organization:

The Comparing Fresno Against Other Geographies plots compare Fresno to other peer cities and counties. CVCF staff members identified other relevant cities and counties both within and outside of California. The goal of these selections was to provide greater context to the Fresno data among other cities and counties in the Central Valley, across California, and among peer cities throughout the United States.

Some metrics are broken down by race and ethnicity or, more infrequently, disaggregated by other variables. The dashboard was built to show the highest degree of racial and ethnic disaggregation possible. If a metric does not show data for specific races or ethnicities, that is because the source data also lacks that racial and ethnic disaggregation.

Departures from the Mobility Metrics and the Upward Mobility Framework:

As noted above, much of the data from the dashboard came directly from the Urban Institute’s Upward Mobility project. While we followed that project’s framework of the predictors and metrics for the majority of the data, we made some adjustments to the data to meet the needs of this dashboard. This section describes those changes and explains why those decisions were made:

We did not include the “Descriptive Representation” predictor. The Planning Guide for Local Action (p27) encourages data users to identify the race and ethnicity of elected officials for this predictor. In the absence of self-reported race and ethnicity information from Fresno elected officials, we chose to report the demographic details of Fresno in greater detail in the Demographic Breakdown tab and direct users to Fresno government webpages to learn more about their elected representatives. See the Source(s) and Notes drop-down for more details.
In the Upward Mobility Framework, the metric used is “the ratio of the share of a community’s home values held by a racial or ethnic group to the share of households of the same group.” Instead of showing this as a ratio of two numbers (share of home values and share of households of a particular racial or ethnic group), we found that taking the difference between these two values was more intuitive. This becomes a one-number summary and therefore no longer reflects the raw shares of home values and households, but we found that the difference is more intuitive and easier to visualize.
To respond to the desire to examine within-Fresno variation in outcomes, we provide tract-level data related to some of the Mobility Metrics. In most cases, the exact metric was not available at the tract level, and we were able to identify a related metric available at the tract level. In such cases, we report the related metric and explain the difference in the variable definitions in the notes and references.
Urban’s Upward Mobility Framework is oriented around five pillars. DRIVE’s portfolio includes three buckets: Prosperity, People, and Place. Consequently, this dashboard shows the predictors and metrics organized by those buckets.
The Mobility Metrics do not include state-wide estimates. All state-wide estimates were calculated specifically for this project. See below for more details.
We incorporate data from four sources outside of Urban’s Upward Mobility Framework as metrics: performance on standardized testing in English Language Arts and Mathematics from the California Department of Education, a water contamination index from CalEnviroScreen, the rate of juvenile arrests from the California Department of Justice, and the share of pre-term births from the California Department of Public Health.

California Estimates:

For data collected independent of Urban’s Mobility Metrics, we collected data both at the county and state scales. To determine a reasonable estimate for Urban’s metrics at the State of California scale, we took a weighted average of the county-level data for each county in California. We chose one of four weights for the data: total population by race and ethnicity, number of households by race and ethnicity, total population under 18 years old by race and ethnicity, and share of voting-age citizens. The expression below shows how we took the weighted average, where \(i\) refers to a specific county and \(j\) refers to a specific mobility metric and its corresponding ACS variable used for weighting.

\[\begin{equation} \frac{\sum_{County_i}^\text{All Counties} Metric_j \cdot ACS_j} {\sum_{County_i}^\text{All Counties} ACS_j}\end{equation}\]

As an example, consider the employment-to-population ratio for adults 25-54. For this metric, for each racial or ethnic subgroup, we used the total population of that subgroup as the weighting variable. See the GitHub README for this project to see more details about which weights we used for each metric. To be transparent, these are estimates because the weights do not perfectly correspond to the metric. In the employment-to-population ratio case, the perfect weight likely would be adults ages 25-54, but we mimic this with the total population which is likely quite proportionately similar.

Data Quality:

The Upward Mobility Framework metrics provide a data quality indicator with three values:

“Strong” indicates the metric is measured with adequate accuracy and sample size.
“Marginal” indicates that there are known shortcomings of the data for this metric, and the metric should be used with caution.
“Weak” indicates that although the metric could be computed, Urban researchers have serious concerns about how accurately it is measured for this community. They recommend seeking more local data sources for this metric.

The majority of the Mobility Metrics shown in this dashboard are “Strong.” However, some data shown in the dashboard are considered “Marginal” or “Weak.” We encourage downloading and exploring data to assess data quality.

Missing Data:

For some measures, some racial groups and/or years of data will not be shown in the dashboard. There are several reasons for this omission:

Missing underlying data: Metrics for some years and/or subgroups may not be available from the original data source. For example, the public transit use and transportation cost measures are derived from data originally from the Center for Neighborhood Technology’s Housing and Transportation Affordability Index, which only releases data for specific years (2015, 2019, and 2022), so this dashboard cannot show data from other years.
Insufficient sample size for disaggregation: Some metrics do not include racial or ethnic subgroup breakdowns because the original data source does not provide that level of disaggregation. The dashboard was built to show the highest degree of racial and ethnic disaggregation possible. If a metric does not show data for specific races or ethnicities, that is because the source data also lacks that detail.
Geographic limitations: Some supplemental data sources (California Department of Education, California Department of Public Health, CalEnviroScreen, and California Department of Justice) are specific to California, so they do not include data for the out-of-state peer counties and cities shown elsewhere in the dashboard.

Data Dictionary for Data Downloads:

Data downloaded from the Download Data (CSV) button will have columns of the following form, depending on the chart and data source:

Mobility Metrics data (Prosperity, People, and Place sections) will have columns of the following form:

year: (numeric) The year corresponding to the data. Note that the meaning of year changes depending on the variable. For some metrics, the year refers to a specific calendar year; for 5-year American Community Survey data, we label the 5-year estimates by their middle year (for example, 2017-2021 5-year estimates are labeled as 2019 in line charts). For more details on the specific meaning of the year, see the Source and Note(s) drop-down for each chart.

state: (character) The state FIPS code.

state_name: (character) The name of the state.

county_name or place_name: (character) The name of the region (county or city) corresponding to the data. For city data, the column name will be place_name.

county or place: (character) The FIPS code for the county or place corresponding to the data. For city data, the column name will be place.

subgroup: (character) The subgroup within the county or city to which the data refers. Common subgroup types in the data are race and ethnicity (e.g., “Black, Non-Hispanic,” “Hispanic,” “White, Non-Hispanic,” “Other Races and Ethnicities”) and neighborhood racial composition (e.g., “White neighborhoods,” “Mixed neighborhoods,” “Neighborhoods of color”). For charts that do not show disaggregation by subgroup, this column will be absent. Neighborhood racial composition groups are defined by their share of non-Hispanic white residents: White neighborhoods have at least 60 percent non-Hispanic white residents, mixed neighborhoods have between 40 percent and 60 percent, and neighborhoods of color have between 0 and 40 percent.

[metric name]: (numeric) The estimate for the selected metric (e.g., ratio_living_wage for Access to jobs paying a living wage, share_employed for the Employment-to-population ratio for adults ages 25 to 54, or pctl_income_20 for Household income at the 20th percentile).

low and high: (numeric) The minimum and maximum values of the metric across the geographies or subgroups shown in the chart. These are used for chart formatting.

[metric]_quality: (character) The strength of the data, as assessed by the Mobility Metrics team. Values are “Strong,” “Marginal,” or “Weak.” See the Data Quality section above for more details.

Supplemental data from sources outside of the Mobility Metrics (California Department of Education, CalEnviroScreen, California Department of Public Health, California Department of Justice) follow a similar column structure (year, state, state_name, county, county_name, subgroup, low, high) with metric-specific columns:

California Department of Education data includes pct_passed_math and pct_passed_ela columns representing the percentage of students meeting or exceeding standards.
California Department of Public Health data includes rate_preterm_birth, rate_preterm_birth_lb, and rate_preterm_birth_ub columns.
CalEnviroScreen water quality data includes a pctle_drinking_water column representing a percentile-ranked drinking water contaminants index.

Demographic data (Fresno Demographics section) will have columns of the following form:

GEOID: (character) The census GEOID for the geography (e.g., county or city FIPS code).

NAME: (character) The name of the geography (e.g., “Fresno County, California”).

variable: (character) The race or ethnicity group name (e.g., “White, NH,” “Black, NH,” “Hispanic or Latino,” “Asian, NH,” “Two or More Races, NH”).

estimate: (numeric) The population count for the given group.

moe: (numeric) The margin of error associated with the estimate.

year_x_axis: (numeric) The year used for plotting on the x-axis.

year: (numeric) The calendar year of the data.

conf_int_bottom and conf_int_top: (numeric) The lower and upper bounds of the confidence interval derived from the margin of error.

cv: (numeric) The coefficient of variation, a measure of estimate reliability.

Inclusive Recovery Indices data will have columns of the following form:

place: (character) The name of the city (e.g., “Fresno”).

state: (character) The two-letter state abbreviation (e.g., “CA”).

metric: (character) The specific inclusion index. Values are “economic_inclusion_index,” “racial_inclusion_index,” or “overall_inclusion_index.”

year: (numeric) The year of the data (e.g., 1980, 1990, 2000, 2013, 2016, 2020).

rank: (numeric) The city’s rank among the cities in the data for the given metric and year.

score: (numeric) The index score for the given metric and year.

The full Mobility Metrics datasets are also available from the Urban Institute Data Catalog. In the full datasets, the county-scale data show mobility metrics for every county in the United States, and the city-scale data contain metrics for approximately 480 cities with populations of greater than 75,000. For this dashboard, we use the overall, race/ethnicity, and race share datasets at both the county and city scales.

We calculated weighted estimates for California specifically for this dashboard. Those California estimate variables are available via the data download buttons below the charts.

Suggested Citation:

Urban Institute. 2025. Mobility Metrics Data for the Upward Mobility Framework (v2025.01). Accessible from https://datacatalog.urban.org/dataset/mobility-metrics-data-upward-mobility-framework. Data originally sourced from, developed at the Urban Institute, and made available under the ODC-BY 1.0 Attribution License.

Updates:

This dashboard has been repeatedly updated to incorporate new charts and updated data. The most recent update was on October 3, 2025.

Other Resources:

To learn more about data privacy and data equity and to see an example of different statistical disclosure techniques, see this dashboard developed by Urban Institute data scientists.

In conjunction with researchers and technical assistance experts at the Urban Institute, eight city or county governments put together Mobility Action Plans using the Mobility Metrics. They can serve as examples of how local governments use these data to increase upward mobility and racial equity.

This dashboard was built in Quarto using the R programming language. To learn more about Quarto, you can read its extensive documentation page. The best place to learn more about R is R for Data Science (2e), though there are many excellent resources.

This dashboard is a collaboration between the Central Valley Community Foundation and the Urban Institute.