Meet the former U.S. data chief taking the ‘Lorax’ approach to disappearing federal data

When asked recently what she does for work, former U.S. Chief Data Scientist Denice Ross came up with a simple explanation that helps her avoid going into too much detail during small talk. “I’m sort of like the Lorax, but for federal data,” she says.

Much like Dr. Seuss’s children’s book character the Lorax, who “speaks for the trees” amid deforestation, Ross aims to speak for federal data through her leadership of the Federation of American Scientists’ budding Data Policy Institute. The analogy is silly, she admits, but her meaning is clear: Federal data might not be as accessible as other issues and stands to benefit from an advocate.

“Everything that we do daily has some sort of connection to federal data,” Ross said recently on the DailyScoop podcast. The institute has been “trying to make that more explicit and visible, so that we’re not taking it for granted.”

The Data Policy Institute launched earlier this year as an arm of the nonprofit science think tank dedicated specifically to advancing the field of data policy. It looks to achieve that through monitoring federal datasets, identifying opportunities for public engagement, and pinpointing areas for improvement within the patchwork environment of federal data policy. 

While the overarching aims of the institute are somewhat evergreen, its creation came during and was shaped by a period of upheaval for federal data. 

Under President Donald Trump, dozens of datasets have vanished and hundreds of categories have been removed or altered, as chronicled in the institute’s recently launched Federal Data Terminations Tracker. In many cases, those changes happened quietly, targeting data that doesn’t align with the administration’s political aims, and have been exacerbated by losses in federal capacity — which Ross said might be the most difficult impact to recover from.

“My hope is that people start to see that invisible infrastructure that is federal data,” Ross said.

Currently, the institute is documenting those changes, but as opportunities arise, Ross said she wants to get people to shift from using the data to advocating for it. So far, she said they’ve found success on that front, getting hundreds to show up to webinars on notices about certain datasets in the U.S. Federal Register. 

The organization’s efforts also coincide with the federal government’s ongoing work to implement the bipartisan OPEN Government Data Act, which will likely be a future focus for the group. That statute, which became law in 2019, established much of the existing management framework in government, including the role of agency chief data officers and their interagency council.

Ross’s hope is that over the next few years, the organization can help build “civic infrastructure,” so that as “the chief data officers start to engage with the public more, that we’re ready.” 

Finding the field

During the Biden administration, Ross became just the second person to hold the role of U.S. chief data scientist. That role, which was established under the Obama administration, can change in focus depending on leadership, she explained. Her aim was to analyze federal data to figure out how policies and programs were impacting the American public.

“I had the best job ever as the chief data scientist because my job was to build the capacity of federal agencies to disaggregate data so that they can better identify communities that might be left behind or overburdened, and then take action to remedy that,” Ross said. 

That’s when data policy as a field became clearer to her. 

“I’ve been in this field for 25 years and didn’t realize until I was in the White House for the second time that I was a data policy person, because it’s not a very well defined field,” said Ross, who also served as a senior advisor in the Obama White House following a tour as a Presidential Innovation Fellow. 

According to Ross, it’s possible to study public policy and never learn about the frameworks that govern federal data, and many data scientists working in the federal government similarly get exposed to those policies while they’re on the job. 

As such, part of the institute’s focus “is formalizing and really advancing the field of data policy so that it can better meet the needs of the American people,” she said.

Ross’s work in that area began before the official launch of the Data Policy Institute.

In May, Ross co-authored a Federal Data Field Guide with Christopher Marcum, a former senior statistician and senior scientist in the White House Office of Management and Budget. Ross said the intention there was to create a “sort of a playful way of organizing the federal data ecosystem.”

That document was published with support from a UC Berkeley technology policy fellowship and outlines categories of federal data through the theme of a nature field guide, complete with animals and bugs as mascots to represent each data type. 

Ross said that the guide serves as a foundation for thinking beyond just statistical and administrative data and is the starting place for data policy education. Already, she used it as the curriculum when teaching congressional staffers that were part of a Data Skills for Congress group this summer.

Tracking terminations

But simultaneously, a new demand for data monitoring has arisen as the second Trump administration has targeted federal data and altered its surrounding infrastructure. That’s where the institute’s data-tracking component comes in.

In the first hours of the current administration, Trump signed executive orders to rid the federal government of programs or policies related to diversity, equity and inclusion and swap out the terms “gender” and “gender identity” for “sex.” Soon after, thousands of datasets were removed across the government. 

While most were scrubbed and reappeared, Ross said, there were losses to gender identity, race, ethnicity, and climate data.

Around that time, efforts to preserve that information began to appear. The Data Rescue Project, for example, began its work to crowdsource an archive of datasets captured before they’re removed or modified. The Impact Project began collecting data on the impact of policy changes within the federal government, and later that year, Ross published the first iteration of what would become the Terminations Tracker. 

On Oct. 31, 2025, Ross and collaborators released a collection of “Dearly Departed Datasets,” which included tools, variables, and full data publications that had been removed since inauguration day. The aim with that project, Ross said, was to create a method for defining a dataset to better quantify the real impact after initial estimates overstated the effects.

“It was our first attempt to crowdsource what datasets had been removed or terminated, and that was really in response to some of those big numbers that were floating around about data losses,” Ross said.

The high estimates for data losses early in the administration were mostly due to people overinterpreting changes to the overall number of datasets in platforms such as data.gov and not having a clear definition of the difference between products, websites, and tools, she said. 

Eventually, Ross and her team realized that there would be benefit from a formalized and ongoing process of tracking federal data, and the Terminations Tracker was launched in August. Today, that tool contains 38 terminated datasets and 337 data elements that have been removed since January 2025.

“Building on Dearly Departed, our goal with this tracker and what we are putting out is to make it the most policy-relevant version of data terminations that we could get,” said Melanie Klein, senior associate of federal data policy at the Federation of American Scientists.

To achieve that, it focuses on two areas, Klein said: The first is whether the terminations are above a baseline of the churn that is expected across administrations, and the second is looking at whether the data collections, models and tools are unique to the federal government — meaning that only the government could fix or produce the data.

So far, Klein said forms and surveys are the easiest to track given the visibility of those changes in the Federal Register and Reginfo.gov. Some agencies will also announce terminations. But outside those methods, the project gets the rest of its information from crowdsourcing.  

“We have search agents that are doing daily searches to find things on news and social media and federal sites, but at the end of the day, we’re going to need to lean on our community of data experts so they can flag for us what datasets are no longer being collected,” Klein said.

Ongoing work

For members of the data-policy community, the institute brings new focus to the space. 

Nick Hart, CEO and president of the nonpartisan, nonprofit Data Foundation, told FedScoop he welcomed the efforts to bring more people into data policy and engage with the public. 

“The national data ecosystem has been under-resourced for years, across administrations of both parties, and it needs far more capacity than any single organization can provide,” Hart said. 

On the monitoring effort, in particular, Hart said even when retiring or altering a dataset is part of the normal lifecycle, “the people who rely on that data deserve notice and a voice in the decision, as the first Trump Administration’s Federal Data Strategy called for.”

Erie Meyer, a co-founder of the U.S. Digital Service and former chief technologist at the Consumer Financial Protection Bureau, said that with so much turmoil — from the war in Iran to the rising cost of gas — it’s important to have a better understanding of what’s being hidden. 

“I think Denise and the other people who are trying to document and catalog what is happening are really critical to our recovery,” Meyer said. 

In addition to the Terminations Tracker, the Data Policy Institute also produces its data health check up via DataIndex.us and has recently been engaged in advocacy surrounding the comment period for sweeping proposed changes to the decennial census. Those changes would, among other things, remove race and ethnicity questions.

Going forward, Ross said she wants to expand the visibility the organization has into difficult-to-track areas, such as satellite and sensor data, and to continue bringing the public up to speed on data issues. Eventually, Ross said, she’d like to get data-policy experts in government.

“I would love to see, in whatever the next administration is, to see data-policy people who are intentionally placed in some of those agencies that have a lot of data issues going on, so that they can hit the ground running,” Ross said.

Serious message

The creativity in the data policy space is evident. Beyond the Lorax comparison and federal data field guide, guests at the launch party for the Data Policy Institute in June were invited to dress as their favorite federal datasets — one even prepared a poem. 

Ross herself was dressed as the North American Bat Monitoring Program, which, as she also explains in a social media video, is a database that provides information about bat populations to better protect them. That’s important, she notes, as bats save farmers billions in pest reduction. 

“One of my biggest regrets from my time as the chief data scientist of the United States was that … we didn’t do a good enough job about telling the story about why federal data matter to everyday Americans, and that storytelling is just really fun,” Ross said. 

But behind the costumes and children’s book analogies that help communicate the importance of data, there is undeniable seriousness to the message that data-policy advocates like Ross are trying to send.

When asked what she hopes the public-sector community knows when it comes to government data, Ross warned that people’s lives will decrease in quality as federal data weakens.

“All federal data right now are at risk because of the collateral damage from the reduction in federal capacity writ large, the loss of staffing, the loss of contractors, the dissolution of the federal advisory committees that help keep data collections relevant and methodologically sound,”  Ross said. “And we should be very afraid.”