Rejoining Federal Service as the next U.S. Chief Data Scientist

[The author, dressed in office clothes, and her young son measuring anti-slip tape for a wooden ramp]
Aug 21, 2006 | My oldest son and I measuring the anti-slip tape for the ramp to his new child care center — a licensing requirement before the center opened a few weeks later

I’m excited and honored to be joining the White House Office of Science and Technology Policy in the role of U.S. Chief Data Scientist. This professional transition — and the times we are in — have me reflecting back on the months and years after Hurricane Katrina and how that experience shaped the work I do.

Many lives were destabilized due to the storm. In my case, two trees fell on our house. We lost power for six months, but didn’t get any flooding. More broadly, the city was reeling from the infrastructure failures of the federal levee system. At the same time, the faltering infrastructure of schools, utilities, child care, medical and mental health care loomed large in our collective attempts to rebuild in a way that included everyone, equitably.

My husband was deployed overseas with the military when our toddler and I moved back home, so it was just me and the kid in a house with no power. The neighbors were kind enough to string an extension cord through the window so I could power a space heater and a dorm fridge. Kind neighbors and clever hacks were a constant theme in the region’s recovery. I am reminded of what historian Andy Horowitz wrote after Hurricane Ida this summer: that resilience “usually means that we should attempt to find individual solutions to our structural problems.”

We sure had plenty of structural problems.

The loss of our child care was a confluence of structural failures. Our son’s child care center took damage in the storm and the owner wasn’t able to move back because her children’s elementary school hadn’t reopened. Then, with property values skyrocketing, the center’s landlord decided to sell the building. To top it off, teacher wages were insufficient to live in this newly overpriced city. At the time, I didn’t fully understand how interconnected these things were, but I felt the lack of child care acutely.

Meanwhile, I had work to do. I co-directed the Data Center, a nonprofit that repackaged publicly available data in a way that was relevant and actionable for local community organizations and neighborhood leaders. We’d spent the years before Katrina trying to infuse data into local decision-making, and now people were begging us for data where there was none.

Indeed, the data we’d been using was rendered instantly historical by the storm. This was all before “open data” — four years before President Obama’s Open Government Directive — so our work was scrappy and involved lots of spreadsheets, DVDs, and scraping public web sites. We started rounding up data that would be relevant to the recovery. Where were people moving back and what services did they need? Which services were open? We catalogued clinics, emergency rooms, police and fire stations, schools, and of course, child care centers.

During the day, my colleagues and I were mapping closed child care centers; on nights and weekends, I volunteered with a group of parents trying to open a new center for our own little ones. In the meantime, we pooled babysitting and took in each other’s children to form a patchwork quilt of care. One year after the storm, the new center opened.

Fast forward to today, where 3 in 10 adults experience child care disruptions. The list of reasons is long. Covid shuttered many centers. Those that remain are experiencing widespread worker shortages, in part because of low wages. At the core, the pandemic has exposed just how brittle our nation’s caregiving infrastructure is.

Brittle infrastructure is the lived experience of millions of families. But with the pandemic, the Federal Statistical System kicked in with timely data to quantify the problem with the Census Bureau’s Household and Small Business Pulse Surveys. The Pulse surveys are just one of many data innovations at the national and local level to come from this crisis.

I’ve written before that if public interest data science doesn’t start out personal, it eventually becomes so. With the pandemic, all of America has experienced living through a disaster and understands how critical data are when the world turns upside down.

The current moment is an extraordinary policy window for the Biden-Harris Administration to shore up our nation’s data infrastructure and reconfigure the public decision-making process so data scientists can best support the design and implementation of policies to improve the lives of all Americans.

Having spent the last twenty years of my career on the boundary between data and action, every time there is a disaster, it still feels like we start over from scratch. Compounding crises continue to pummel our communities in the form of wildfires, extreme weather, infrastructure failures, police violence, hate crimes, and civil unrest.

This pandemic — similar to Katrina and all disasters — amplified existing disparities across racial, ethnic, income, and gender lines. Though obvious to those who are impacted, these disparities are often invisible to policymakers because of incomplete, inaccessible, inaccurate, and/or lagging data.

But we can get ahead of this. As I head into this new role as Chief Data Scientist, I’m focused on addressing these questions: What data do we need to design an equitable recovery? How will we use that data responsibly and to ensure we are building back better? What can we do to support the Federal, Tribal, state, and local data practitioners so they can help deliver better outcomes for the American people?

The shocks and stressors of the past 18 months represent a threshold event that can dramatically change the role of data science in our society for the better.

With historic levels of resourcing through the American Rescue Plan, Build Back Better, and other Federal investments, we have a unique opportunity to infuse data into policies and programs that will deliver on the Biden-Harris Administration’s priorities of pandemic recovery, the climate crisis, economic prosperity, and racial equity. However, that opportunity is time-limited, racing against a strong societal pull to “get back to normal.”

Our challenge is how to create durable change at scale within the shortest time possible. And we can’t do it alone. As my predecessor and mentor DJ Patil likes to say, “data science is a team sport.”

I come into this role having worked on great teams – most recently, co-authoring the Pandemic to Prosperity series tracking the data necessary for an equitable recovery, and advocating for greater data transparency at the Federal and local level on issues as diverse as the 2020 Censuspolicingdisasters, and climate change. When I co-led the Police Data Initiative in the Obama Administration, I had the privilege of working with police chiefs and data leads from more than 100 law enforcement agencies. During my time at the City of New Orleans, we partnered with Code for America to activate data about blighted, storm-damaged properties to accelerate neighborhood recovery. And from the decade I spent working in the nonprofit sector, we had to get creative with partnerships to fill critical data gaps.

The mission of the U.S. Chief Data Scientist is to responsibly unleash the power of data to benefit all Americans. I’m eager to get started and look forward to collaborating with data innovators already working inside of Federal government and at the State, Tribal, and local level.

Please follow me on Twitter @DeniceRoss46 and stay tuned for ways to scrub in. We’ve got a lot of work to do!

5 Data Lessons from the Pandemic

By Denice Ross and Allison Plyer

Eighteen months since the Covid pandemic upended the world, and sixteen years since we learned this lesson from Hurricane Katrina, good data is the path toward an equitable recovery.

Here are 5 lessons learned from the pandemic and our careers in data and disasters:

1/ Design data to be granular enough to show which demographics and geographies are disproportionately impacted.

2/ Invest in data capacity now—to help with the current crisis and also with future shocks.

3/ Address data quality, standards, and timeliness to shed a more complete light on rapidly changing conditions and the needs of the most vulnerable.

4/ Crises exist in context. Incorporate a broad range of data into decision-making to ensure critical aspects of our society and democracy don’t get derailed. 

5/ Build partnerships and support community organizing to turn data into action. 

Read the full article here.

Pandemic to Prosperity

History has shown that large-scale crises accelerate pre-existing trends, exacerbate inequities, and permanently change societies and civic life. Large-scale disasters produce an enormous break in the status quo followed by continuous change. Recovery from the pandemic and deep economic crisis will vary across communities, and different populations will face various barriers to achieving shared prosperity.

Pandemic to Prosperity offers a comprehensive overview of the Covid-related impacts on our lives and livelihoods, governments, civic institutions, and overall well being. This report series analyzes disparate data, adding top-level insights about the implications of each indicator, what each indicator reveals, and how the indicators are interrelated. It highlights mostly state-level metrics with breakdowns by race, gender, age, and income where available, relying on both public and private data sources.

Reports are archived at, and I co-authored the series from its inception to the Oct 2021 report focused on the South.

Police data belongs to the people

Radical transparency in policing would be an important departure from the status quo. Here are five data sets departments should start sharing widely.

By Clarence Wardell and Denice Ross | Boston Globe June 12, 2020

Houston police chief Art Acevedo, an advocate for public disclosure of data about policing, at a public visitation for George Floyd. ERIC GAY/ASSOCIATED PRESS

Prompted by the recent police killings of George Floyd and Breonna Taylor, protesters are demanding a wide range of changes to policing, including abolition, shifting funds to other community services, and tactical reforms. A common thread across these demands is that American policing must be held accountable to the communities it serves. Accountability, however, requires transparency — and transparency is a concrete step that local leaders can take right now.

After Michael Brown was killed by police in Ferguson, Mo., in 2014, the Obama administration launched the Police Data Initiative as part of a detailed national response to racialized police violence. At the time, few if any police departments in the country published data about their own actions in sufficient detail for community members to check for evidence of bias.

These days, information about police officers’ actions in addition to the arrests they make is more commonly released. But most of these data sets still lack key details and crucial context, such as corresponding body-camera footage, or published policies on what is allowed (and not) when officers use force. In most cases, releasing data isn’t mandated by law; it’s a matter of what leaders want to do. The Boston Police Department, for example, stopped publishing its annual data on stop-and-frisk incidents after 2016. It took months of public calls for transparency, public records requests, and finally a subpoena to restore the flow of data just last month. BPD’s excuse for the three-year gap in publishing data? Nobody had asked for it.

Americans shouldn’t have to beg for data from agencies that have such extraordinary powers. As Art Acevedo, then the police chief in Austin, Texas, and now the chief in Houston made clear five years ago: “This isn’t our data, it’s the people’s data.”

Governments should lean into the idea of being held accountable by their community members in ways that would represent a radical departure from the status quo. It is necessary for both legitimacy and trust. Leaders can start immediately by ordering the release of these five data sets:

  • Use of force, including shootings by officers. Is force more likely to be applied in communities of color, adjusting for other factors? What are the results from internal investigations into whether the force was justified? The Seattle Police Department’s use-of-force data is updated automatically in near real-time, and Orlando’s officer-involved-shooting data includes detailed review letters from the State Attorney for each incident.
  • Complaints against officers. What complaints are people filing about police officers? How are these complaints against officers resolved? The Citizen Complaint Authority in Cincinnati helps the public understand this data in graphs, charts, and maps, making it easier to devise better policies.
  • Police force demographics. Does the police force look like the community it serves? Are they failing to retain women and people of color? Wallkill, N.Y., publishes an annual spreadsheet that details rank, years on the force, gender, and education levels of the 120 people in their department.
  • “Stop-and-frisk.” Which populations are police most often stopping in the field, and for what reasons? The Boston Police Department’s newly liberated data includes the names of the officers making the stops and their supervisors. NYPD releases annual data with demographic details and the reasons for the stops.
  • Traffic stops. Are people of color disproportionately likely to be pulled over? Are police actions biased, whether they let someone off with a warning or ask to search the vehicle? The San Diego Police Department, in accordance with the California Racial and Identity Profiling Act of 2015, releases demographic details on the people stopped, as well as reasons for the stops and any actions taken by the officers.

Numbers alone won’t tell the whole story, though. Radical transparency will require police and other government agencies to publish complementary records and documents, such as the department’s policy handbook (including the rules on the use of force), police union contracts, prosecutorial and review board decisions, and internal disciplinary records. Departments should promptly make body-worn camera footage available when an incident is being reviewed to clarify, for example, whether a person “tripped” or was actually pushed by officers.

We need to also ensure that all data are released responsibly, protecting privacy so that victims of crimes and police misconduct feel comfortable reporting. Greater transparency will also, in some communities, require revisiting outdated laws and obstructionist police union contracts that are holding back data to which the public is entitled. Leadership is essential to breaking these logjams.

The lack of transparency has not only left our law enforcement apparatus unchecked and unaccountable to the community, but it also has made it harder to understand what actually works to reduce police violence. After the death of George Floyd, we learned Derek Chauvin had at least 18 citizen complaints filed against him. Accountability starts with transparency. We must face the difficult truths hiding in the unopened vault of police data.

Clarence Wardell is director of City Solutions for the What Works Cities Initiative at Results for America, a research organization that advises governments. Denice Ross is a fellow at the National Conference on Citizenship and Georgetown’s Beeck Center for Social Impact and Innovation. They co-founded the White House Police Data Initiative in 2015.

The Hill: Transparency is essential for a successful 2020 Census during the COVID crisis


With the Census Bureau slowly re-starting field operations, it’s a good time to take stock of the level of transparency needed to mobilize the nation for a complete count in these times of COVID. 

We have long known that the 2020 Census will require an all-hands-on-deck approach. Counting those who are hardest to reach (people of color, complex households, families with young children, LGBTQ individuals, etc.) requires the Census Bureau to coordinate with organizations these populations trust. With this in mind, the bureau has cultivated a national network of thousands of Complete Count Committees and Partners. These nonprofits, faith organizations, local governments, businesses, and community-based organizations serve as trusted voices to reach communities skeptical of the census. 

I know first-hand why transparency matters during crises. After the levees failed and flooded 80 percent of New Orleans in Hurricane Katrina, the recovery effort was massive. It required the government to lead and bring these same stakeholders together as director of Enterprise Information for the City of New Orleans. 

It was my responsibility to help build a common base of shared information from which to start. We published unprecedented amounts of data about the city’s operations, giving community stakeholders the information they needed to help advance the recovery. Open data became an engine of coordination and community innovation that was essential for revitalizing storm-damaged neighborhoods.

As the United States looks to recover from this pandemic, the census data being collected now will be critical. Historically, more than a third of households don’t self-respond, requiring census takers to collect responses in person if possible. 

However, the quality of the data is better when people respond on their own, and follow-up field operations are expensive-more than half the total estimated cost in 2020. Minimizing field operations is also now a public health imperative. We can shrink the need for in-person interviews if more households fill out the form themselves by phone, internet, or paper.  

COVID-19 has massively disrupted the bureau’s decade-long planning. To their credit, the career professionals are adapting, but so too must their partners. These frontline organizations urgently need more data about the bureau’s operations to make their efforts complement those of the bureau’s, rather than working at cross-purposes. 

For example, local census partners are spreading the word that residents who need help filling out the form can call the Census Bureau. However, if the bureau were to share data on call volumes, then partners could tell residents when the best time to call is, rather than clogging up the phone lines and causing long wait times that discourage residents from completing the process. Complete Count Committees and Census Bureau Partners need that type of data in an open format so they can incorporate it into their own carefully planned outreach efforts. 

The Census Bureau has already taken great strides to turn transparency into collective action. They publish a map of all Complete Count Committees to help volunteers find a local get-out-the-count effort. Their “Recruiting Goals by County” map helped local organizers identify candidates for census jobs. And most visibly, the bureau publishes updated self-response rates daily for areas as small as a neighborhood. A whole ecosystem of users has emerged around this open data, with mayors competing against each other for the highest response rate, and 56,000 visitors so far to CUNY’s open-data-fueled Census Hard to Count Map.  

To counter the disruption brought by the virus, the Census Bureau should double down on its success with data transparency. This way, civil society can align efforts with the agency’s and make sure that all people are counted — safely. 

 What type of data would be most useful?

  • Continuing the daily flow of self-response rates and adding weekly analysis on response rates for specific hard-to-reach populations like children ages 0-5 and renters to hone partner outreach messaging
  • Completion rates for the non-response follow-up workload so trusted local messengers can encourage participation
  • Group quarters workload completed by type of facility (nursing home, college dorm, correctional facility, etc.) and by state, so state and local officials can provide support for sectors falling behind in the count 
  • From the census call centers, hourly wait times by language line, and the most common questions from callers, such as “when will I get my paper form?”
  • If the bureau needs more workers, recruiting goals by county with needed demographic characteristics and language skills to empower local organizations to identify candidates

Government transparency is typically thought of as a tool for accountability. Perhaps its highest purpose, though, is as a tool for enabling the government and civil society to work together to accomplish goals otherwise unachievable. Getting a fair and accurate count for the 2020 Census during a pandemic is exactly that type of challenge.

Denice Ross is a senior fellow with the National Conference on Citizenship and Georgetown’s University Beeck Center for Social Impact + Innovation. She is a former Presidential Innovation Fellow, co-founder of the Police Data Initiative, and launched the City of New Orleans open data initiative.

How Data Became Part of New Orleans’ DNA during the Katrina Recovery

Data intermediaries have a symbiotic relationship with government as the source of most of their information. The open-data movement in government and development of software-as-a-service technologies shaped the data landscape after Katrina. Through relationships and talent transfers with The Data Center, the City of New Orleans went from having its chief technology officer in federal prison and its data systems in shambles to being a nationally recognized leader in open and accountable government. To be effective during disasters, an intermediary should be (1) in place and widely respected before the event, (2) ready to respond immediately after the event and for the long recovery, and (3) continually scanning the horizon for changes in data and technology

Recommended Citation: Gardere, Lamar; Plyer, Allison; and Ross, Denice (2020) “How Data Became Part of New Orleans’ DNA during the Katrina Recovery,” New England Journal of Public Policy: Vol. 32 : Iss. 1, Article 21.

Read the full paper at:

Towards a Deliberate Practice of Public Interest Tech

After nearly twenty years in the field, I’ve learned that if public interest tech doesn’t start out personal, it eventually becomes so. After Hurricane Katrina, as I negotiated with Louisiana state government for access to their childcare database, I was also patching together itinerant care for my own children, since 80% of childcare centers in New Orleans were shuttered. As we compiled data to support the various community planning processes, an illegally placed fast food restaurant popped up across the street from our home. And, while my organization was trying to figure out the storm’s final death tally, I read in the New York Times that our pediatrician had died by his own hand; I had to wonder, should suicide three months later count in that tally?

Read more of this article at New America.

The Phase Zero Digital Toolbox

Visualizing global security, state instability, climate change, and vulnerability of natural resources


People have been using tools to adapt to the environment since the beginning of humankind. In today’s environment, digital tools — not just tools of stone and metal — are increasingly helping humans adapt.

Today, complex interactions between demographic change, rising standards of living in many parts of the world, new technologies, and changing demands on natural resources are raising the risks to global security and prosperity. Climate change, in particular, means unprecedented shifts in the risk factors for conflict, social unrest, and forced migration. Digital tools can help governments, businesses, and other organizations better understand, anticipate, and plan for today’s dynamic human security environment.

The Phase Zero Digital Toolbox is a curated list of existing digital tools for understanding complex interactions. These data visualization and decision suppo rt tools look at root causes of insecurity, such as: water and food scarcity; the record of humanitarian crises, conflicts, and state instability; and impacts of global climate change. Some make integrated assessments of multiple trends and variables. Some are purpose-built for specific end-users. Many are intended to raise public awareness. Most look at a range of data, everything from satellite imagery to on-the-ground surveys, and represent them as graphs or maps.

The Phase Zero team assembled this toolbox primarily to support our own research. We are sharing our research for two reasons: 1) to showcase the wide range of available digital tools in order to spur cross-pollination of approaches, and 2) to help would-be users of these tools find them. For inclusion, a tool must be free and publicly available, have multinational and ideally global coverage, and be published by a credible organization using documented sources. This is intended to be a sampling, not a comprehensive list of tools, and there were more tools behind firewalls that we were not able to evaluate.

Read the full report at New America...

The Police Data Initiative Year of Progress

How We’re Building on the President’s Call to Leverage Open Data to Increase Trust between Police and Citizens

In the fall of 2014, we began our time in the Administration as Presidential Innovation Fellows. Our service took place against the backdrop of a national conversation on police reform and how to increase trust between police departments and the communities they were sworn to protect and serve.

Emerging from that conversation, The President’s Task Force on 21st Century Policing highlighted the opportunity for technology and data to play a central role in building trust between law enforcement and communities while enhancing public safety.

As civic technologists, we had an idea to build on that opportunity:

What if we could help enable a new culture of open data in law enforcement agencies where police collaborate with their tech counterparts in local government and the community to publicly release incident-level, structured, machine-readable data on policing?

Today, the White House is welcoming 39 law enforcement agencies from across the country who are leading the way toward developing a culture of open data in policing through the President’s Police Data Initiative. Joining these agencies are community stakeholders, civic technologists and data scientists, researchers, and local officials.


Read the rest of this article by Clarence Wardell and Denice Ross on the White House Medium channel…

Ten Years After Katrina: New Orleans’ Recovery, and What Data Had to Do with it

map Open data matters most when the stakes are high

As a New Orleanian, it’s hard to believe how far we’ve come in the 10 years since Hurricane Katrina and the levee failures. In the days of the aftermath, we had a tremendous effort ahead of us, and government at all levels did too: Understanding the extent of the flooding, conducting damage assessments, providing temporary housing for displaced residents, rebuilding the levee system, distributing rebuilding dollars, and issuing building and demolition permits. One of the great lessons we learned through the experience was the power of data to illuminate our path to recovery.

Ten years ago, the concept of “open data” had not yet taken hold within the government.

Back then, accessing even basic government data involved a formal public-records request and often came with restrictive data-sharing agreements. As a result, in post-Katrina New Orleans, the public didn’t have easy access to many government data sets tracking recovery activities. The public could view some government records, one at a time, but because the data were not available in their entirety — in a structured, machine-readable, “open” format — citizens couldn’t download, analyze, or innovate on these data sets.

Read the rest of Denice’s article on the White House Medium channel…