Championing Public Data: Data Professionals in Civic Life

The following text is from Lynda Kellam’s opening keynote for the 2026 IASSIST conference on June 2, 2026.

Lynda Kellam’s opening keynote

Good morning, everyone. It is a genuine pleasure to be with you today as your opening keynote speaker. I know that virtual conferences are challenging. I miss the hallway conversations with every ounce of my extroverted being. Recently, I attended another virtual conference where the opening keynote asked people to stay as present as possible, and I ask that of you now. Not just for me but throughout the week.

What we have to talk about matters, and we need each other to be present and connected.

The IASSIST Community

I’ve been in IASSIST since 2007, and along the way, I’ve had the extraordinary privilege of learning from former IASSIST leaders, such as Chuck Humphrey, Bill Block, and Ann Green, as well as long-time members like Walter Giesbrecht (who passed away this year), Jen Green, Tuomas Alaterä, and many more. I’ve gained close friends and collaborators who have become some of the most important people in my professional and personal life. And I have watched this community grow in size, depth, and diversity of topics.

One moment has stayed with me for years. At one conference, Chuck Humphrey, a former President of IASSIST and a mentor to many of us, was asked at the last minute to fill in for a missing keynote speaker. He had nothing prepared. Instead of talking for an hour, he invited the community into a conversation. He asked questions and listened. And people talked about it for years afterward because he brought us together as a community. That is one of my core professional memories. And, as hard as it may be, that connectedness is part of what I want us to try to achieve with each other this week, even across a screen.

We are a close-knit community. And that closeness, I would argue, is exactly what got us through 2025.

IASSIST 2026 Conference Theme

The theme of this conference—Championing Data: Data Professionals at the Nexus—reflects the difficult reality we are in. But it also points to a belief that although we may be up against many challenges (political, financial, technological, and more), we are facing that future together. And that matters.

In this talk, I want to reflect on the role we play in academia and in the wider world. I want to think with you about who we are, what we have as a data community, and what we’re being called to do next. My argument is that our work as data professionals is civic work, even when we describe it as service or support. The events of the past year have made that civic role impossible to ignore.

Data Unicorns

I’ll start by talking about how I got here, because I think my trajectory reflects the larger path of our field in the last few years, at least, and where it’s going.

I started my career as the first data librarian at the University of North Carolina at Greensboro in 2007. In some ways, that was an incredible opportunity. I got to define what the role meant. In other ways, it was genuinely difficult, because nobody else quite understood what I did. If a question at the reference desk had the word “data” in it, it came to me. If a newspaper article mentioned data, I was asked for my thoughts, even if the content was unrelated to my work.

I was a unicorn. I was expected to know statistical software packages (as many as possible), GIS, government data, my subject domains, and, later, data management practices and qualitative data analysis software. And I tell you this not to complain, but because many people in this room know what I am talking about. As a community, we’ve often faced unrealistic expectations or a lack of understanding of our field and exactly what a data librarian or data professional does.

A Field Transforming

Since 2007, the field I entered has transformed quite a bit. This isn’t to say there wasn’t change before, but I’m speaking to my experience and the timeline that has shaped my career.

Right after I started my job, the fallout from the Great Recession began to reshape our libraries. We were promised library retirements that would open the field to new librarians, but instead experienced belt-tightening. While a rough start, there were an incredible number of positive changes, though. RStudio arrived in 2011 and, along with other open-source tools, reshaped our practice in ways we are still building upon. In 2013, the OSTP public-access mandate helped launch an increased emphasis on data management. And closer to my heart, in 2016, Mandy Swygart-Hobaugh and others began recognizing that qualitative data deserved the same level of support as quantitative data, which is how the Qualitative Social Sciences & Humanities Data Interest Group came to be.

Our field has moved from unicorn positions that required a single person to know everything to specializations with diverse, sophisticated expertise. IASSIST has shifted from focusing on data for social scientists to supporting data across every discipline — from the sciences to the humanities to GIS. The FAIR principles have become a global touchstone, and data management and sharing are an assumed reality at least in our circles, with the CARE principles as a response in 2019. DDI products have been codified by an ISO standard. We’ve seen Canada build a model of national collaboration that frankly makes the rest of us a little envious. We’ve seen or contributed to the development of peer organizations in RDAP, the Data Curation Network, and the regional bootcamps, all of which amplify our work. And in 2021, IASSIST, under the leadership of Winny Nekesa, launched the Africa Chapter, which is now in its sixth year.

We’ve also faced incredible challenges that made the Great Recession seem like a minor blip. In 2020, the global pandemic required all of us to pivot in ways none of us imagined. Since 2022–2023, Generative AI has entered the public consciousness and impacts our work in ways we are trying to navigate and comprehend. And then in 2025, we had massive political and financial disruptions. I don’t think anyone would be surprised to hear me say that the moment we are in is different, especially but not only in the United States.

Beyond the enrollment cliff that we knew was coming, we’ve faced something many of us did not anticipate: a convergence of attacks on higher education, on government data infrastructure, and on the public’s right to know. Federal funding for universities has been threatened. Government employees have been removed at a startling pace. Government data and information have been deleted, modified, and made inaccessible.

The Strains

And libraries are asked to respond to all of this while operating with fewer resources and librarians already at capacity. We are all experiencing the reality that is the theme of this conference: the convergence of multiple kinds of instability at the same time.

Higher education is under financial and political strain. Public trust in institutions has been damaged. The information environment is, in some ways, fragmented and, in other ways, homogenizing. Government capacity is shrinking precisely when people need reliable public information the most. And this is happening while the complexity of our work continues to increase. Many of us entered this profession with an assumption that the overall infrastructure of public knowledge would remain relatively stable. We knew that survey questions might change, technologies might evolve, and funding might fluctuate, but the broader systems themselves would continue functioning. I don’t think we can make that assumption at this moment.

We are seeing more than just changes within stable systems. We are seeing instability in our systems and institutions. Many of us have spent the past few years trying to understand why the work feels different. Why does it feel harder to plan, and why do so many people feel exhausted even when they are deeply committed to this work?

Part of it is that we are no longer only responding to the needs of our patrons. We’re also responding to instability in the institutions and infrastructures around us. And librarians, historically, are very good at absorbing institutional instability. We make do in hard times. Our profession exists to help people navigate confusing systems. We are good at creating workflows and processes. But I know many of us are reaching our limits. Because the scale of the instability is larger than what individual librarians or even individual institutions can absorb on their own.

All of that is the context for both me and the wider profession as we entered 2025, which was an overwhelming year. I think that’s an honest description of where many of us were. I often feel the weight of everything happening in our world right now, and I know I am not alone in that. Every day, I was, and still often am, shocked or appalled by the news. It’s hard to live life in a healthy way when you also feel constantly full of adrenaline from the stress. To some, that may seem like an exaggeration, but that was how I felt.

But one of the things I’ve learned over the past year is that the best response to chaos is to focus on what you actually know *how* to do. To hone in on your expertise and act from there. And for librarians, that instinct seems second nature. When information access is unstable or uncertain, we respond. We love to document, organize, and preserve. We try to create pathways through confusion. Part of why the Data Rescue Project moved so quickly is that many of us recognized the situation immediately, even if we did not understand its scale. We knew what it felt like to see information becoming harder to access. And we knew that if we waited too long, we might lose more than we could recover.

Data Rescue Project as Community Action

Many of you have attended our DRP presentations, many of which I have given. I realize that some of this story is familiar. But for everyone’s benefit, I’ll give a brief version. During 2025 in the United States, we saw a convergence of shifting political priorities, ideologically driven restructuring, budget cuts, and administrative changes that together affected how federal data was collected, maintained, and made accessible.

Members of IASSIST, RDAP, and the Data Curation Network created The Data Rescue Project in early February 2025 in response to the administration’s removal of CDC data and broader disruptions across federal agencies. We were also seeing the loss of significant expertise across federal agencies, making data infrastructure increasingly vulnerable. Because, as we all know, data depends on the people who steward it.

We became a grassroots volunteer network organizing rapid-response rescue efforts when federal data are at risk. We also coordinate preservation and access, and build capacity so that more people know how to act when data are threatened.

And I want to reiterate that IASSIST was a key part of the DRP’s creation. DRP exists because of the relationships and values that IASSIST members have been cultivating.

On February 3, 2025, many of you may remember the flurry of emails on our IASSIST listserv asking about alternative sources for the CDC data that had been removed. Especially the Behavioral Risk Factor Surveillance System, a key resource for understanding health risks. In response, I created a guide that included both alternative resources and existing data rescue organizations. Instead of a LibGuide, I created a Google Doc because I wasn’t sure at the time whether I should or could use my institutional infrastructure for that work.

That Google Doc went viral. It spread across librarian networks, research communities, and the Data Hoarders Reddit. People started crowdsourcing it. They added their own efforts, their own organizations. We learned about some of our closest partner organizations through that Google Doc. We got to know many people across the country and in other countries who were concerned about public data.

What struck me at the time was not the scale of the response but its speed. People were clearly looking for a way to help. They did not need to be convinced that this work mattered. They needed somewhere to direct their energy and expertise. And that was one of the first moments where I realized that the data community had already built the foundation for this kind of response. The relationships and trust in our community already existed.

We had a commitment to the cause. We needed coordination. Those of us in the data librarian community knew people were willing to act, but we needed a way into action that felt responsible and coordinated. We were able to create that path quickly because of the relationships the data community had already built. We brought in members of Saving Ukrainian Cultural Heritage Online, the Environmental Data & Governance Initiative, and the End of Term Web Archive, who had similar or related work and could guide us in launching a broader data rescue effort.

We also learned from the 2017 data rescue efforts. You can learn more about those earlier efforts in Laura Rothfritz’s paper later today, Anticipatory Maintenance and Emergency Curation. While Laura’s work focuses on the 2017 period, many of the challenges she highlights are either ones we have learned from or ones we still face today.

This has not been a perfectly smooth or centralized process. It was messy at times. Volunteers were working across time zones and cultures. We were building workflows while simultaneously responding to urgent requests. We were figuring out questions of ethics, duplication, metadata standards, preservation priorities, and volunteer coordination in real time. But people kept showing up at our office hours and in our communication channels. They shared expertise and connected us to others who could help. And over time, what felt like a loose collection of volunteers began to function like actual community infrastructure.

The DRP Steering Committee — Lena Bohman, Kathleen Burlingame, Halle Burns, Tess Grynoch, Sebastian Majstorovic, Mikala Narlock, Amy Nurnberger, and myself — emerged from the initial discussions. And the community that gathered around us has grown to include over 900 librarians, archivists, activists, technologists, researchers, journalists, and members of the public. Thousands more follow our work through social media and our newsletter. And in many cases, the volunteers are librarians, doing what librarians are trained to do, at a civic scale and under pressure.

That phrase (“at a civic scale”) is key because I think it captures something important about our current world. The actual skills involved in data rescue are not new to us. Metadata creation, workflows, and documentation are not new. Coordinating access is not new. What changed for us was the frame. Skills that many of us practiced within institutions have become visibly connected to public life and democratic access to information. The DRP is evidence of what this community can do when it acts on its own values.

We now know more about the real impact of data loss. Over the past year, roughly 3,000 datasets have been fully removed from federal sources, many of those from USAID. That number may sound small, but behind each dataset is a tremendous amount of labor. In addition, we know that the Department of Agriculture terminated a report on household food security. The CDC stopped releasing data on maternal and infant mortality and reduced staffing for the Pregnancy Risk Assessment Monitoring System. Gender identity questions were removed from the National Crime Victimization Survey. The Department of Homeland Security ended access to HIFLD, a geospatial dataset essential for regional planning.

As data librarians and stewards, we are all aware of why this loss matters. But I want to make it concrete, because I think we sometimes talk about data loss in ways that only resonate with people who already understand what is at stake. Here are two user stories we received:

We heard from a social and psychiatric epidemiologist who focuses on improving the surveillance of eating disorders among youth using the Youth Risk Behavior Surveys. They noted that the prevalence is high, especially among youth of color and LGBTQ+ youth. Youth Risk Behavior Surveys are the only source of updated, representative data on eating disorders among US youth. The CDC’s decision to not process gender identity data means that her team won’t be able to obtain estimates of eating disorders among trans youth, who experience particularly high rates of eating disorders.

Another data user described using the Integrated Postsecondary Education Data System (IPEDS) to help students navigate the college application process and find colleges that fit their particular needs. Without that data, finding appropriate higher education options would be “much harder, possibly impossible, to source.”

These are not hypothetical scenarios. Those are from real data users.

And their stories helped to change the way I think about my work. Once you hear from people whose work and communities depend on this information, it becomes difficult to think about data work as only a research service. Public data is part of the social infrastructure people rely on to understand the world around them. And when that infrastructure becomes unstable, the consequences have impacts far beyond universities.

To back up our rescues, we partnered with ICPSR at the University of Michigan, which hosts DataLumos—a repository created specifically for government data that provides persistent identifiers, metadata fields, and long-term stewardship. ICPSR was one of the first institutions to publicly and courageously support DRP’s work in the early days of 2025. Without Maggie Levenstein and ICPSR’s willingness to let us use DataLumos, none of our work would have been possible.

And I want to emphasize this point (as a slight segue): institutional support matters. Individual volunteers can accomplish extraordinary things, but we need institutions willing to step forward publicly and take responsibility for sustaining infrastructure. One of the lessons of the past year is that we need both: grassroots action and institutional commitment. Neither is sufficient on its own. Institutions need to have as much courage as the individuals doing the work.

And our volunteers have done an incredible amount of work. Over the past year, thanks to them, we have preserved more than 2,500 at-risk datasets from more than 90 federal agencies. We have hosted community events and hackathons, presented at conferences around the world, and been covered in Le Monde, The Guardian, The New Yorker, and many other outlets.

And this work continues. We are not done. We still have requests for our volunteers to back up data.

We are also thinking about the future. We know that data infrastructure is fragile, and that what happened in the United States in 2025 is not only an American problem — it’s a problem that can emerge anywhere. Many of us understand that what we are building is not just an emergency response. We want a broader conversation about what resilient public data infrastructure should look like in the future. We are asking lots of questions in partnership with other organizations, including the Center for Open Science, Invest in Open Infrastructure, and the American Geophysical Union. These questions include: What kinds of networks do we need? What kinds of partnerships? What responsibilities belong to governments, and what responsibilities belong to NGOs, states, libraries, and public institutions when governments fail to maintain access?

In addition to working within those broader efforts, the DRP wants to leave behind documentation for the next generation of data rescuers. You can hear more about that work in Session G Block, From Emergency to Infrastructure: Building A Data Rescue Toolkit, which Halle Burns, Mikala Narlock, and I will facilitate later this week.

But here is what I want you to take from the DRP story: DRP did not create the community that made it possible. It drew from and enhanced that community. The community was already in this organization, in these relationships, in the professional values we have been building together for decades. What 2025 gave us was a moment urgent enough to make all of that necessary. Just as I learned from my mentors in the data community, I want others to learn from what we’ve built. That is how this community works.

From Campus to Civic Life

And this brings me to my larger point today. When data librarians talk about our work, we tend to talk about our patrons and our institutions. And that makes sense. Our libraries exist to serve those communities, and we should be proud of how well we do it. But I want to push on that idea a little. Data curation, data sharing, and helping people navigate complex information are skills that matter beyond our campuses. They matter to people who will never set foot in a university library. And the events of the past year have made that clearer to me than ever.

The rise of AI adds another layer of complexity to all of this. I know there are differing perspectives on AI in this community. But one thing I think we can agree on is that AI has accelerated the expectation that information should be immediate, frictionlessly accessible, and endlessly available.

At the same time, it has increased the need for mediation and expertise. People come to us because they need help understanding where the data came from. Or because they are trying to determine whether something is reproducible, ethically sourced, or even real. So while there is often a public narrative that AI will reduce the need for expertise, many of us are experiencing the opposite. The questions are becoming more complex. And the human labor required to help people navigate that complexity is becoming more important, not less. But institutions are not necessarily resourcing that labor appropriately. And that creates a real tension in this moment. Many of us are being asked to operate as technical experts, educators, policy interpreters, in addition to being data librarians. That is a tremendous amount to ask of a profession that is already stretched thin.

And in an era increasingly shaped by AI-generated information, our profession’s long-standing principles become even more important. The work we do is not peripheral. Our skills are central to how people determine what information they can trust. That matters to people far beyond universities, and it leads directly to what I want to say about who we are really serving.

Think about who depends on federal public data. This slide has just a few examples (in addition to the ones I’ve already mentioned). We do data rescue work; we are fighting for our researchers, but we are also fighting for everyone else who depends on public data: journalists, community organizers, city planners, and parents.

Our work as data librarians supports the broader democratic infrastructure. Access to data is access to information, and access to information is how people participate in public life. The Data Rescue Project and our friends at Data Index, EDGI, PEDP, and other groups have made that clear. Our skills have always been civic, and our communities need us.

In the early days of the Data Rescue Project, journalists would call us the resistance. And I would push back on that framing. In those interviews, I would say our only goal was preservation. I was cautious about the political characterization. At the time, I was protecting my institution, and I understood why that caution was necessary. But I have had time to think about it. And those journalists were right, and I was telling only part of the truth.

My goal was never *only* preservation. My goal — our goal — is access, which in the current environment becomes a political question. In a democracy, people have the right to access public information. The federal government is the only institution that can collect certain kinds of public data at scale, and it is the only institution mandated to share that information with the public without charge. When that data disappears or becomes inaccessible, something essential to democratic participation is lost.

What does that mean for how we work and how we talk about our work? The skills are the same. I am asking us to consider where we show up.

I’m not arguing that every data librarian needs to become a public spokesperson or policy advocate. People have different institutional contexts and capacities. I understand that. And I am also not arguing that civic engagement is somehow separate from librarianship. In many ways, I think it has always been embedded in the work, even when we described it in narrower institutional terms. What changes is our understanding of who ultimately benefits from those skills and what is at stake when that work becomes impossible.

Many of the things we do as data librarians may feel like technical or service-oriented tasks inside our institutions. Many of us frame our work primarily through our campus, our library, our faculty, our students. And those commitments remain essential. But outside those institutions, the skills we have become part of the infrastructure people rely on to understand public life. Our expertise operates within a much broader information ecosystem shaped by political instability, technological change, declining public trust, and growing uncertainty about the durability of public information itself. That means the choices we make about access, preservation, documentation, transparency, and stewardship have broad implications. The past year has made that visible. That visibility changes how we should think about ourselves as a professional community.

These civic roles include activities we take for granted:

• documenting changes to datasets
• teaching students that public information systems are vulnerable to political and institutional change.
• preserving local copies of data and documentation
• providing public comment when surveys or methods are threatened.

The folks at Data Index have done excellent work, especially by surfacing opportunities to comment on changes in federal surveys. And I’ve seen people in our own community begin stepping into that space more often. My own supervisee recently told me how meaningful it felt to submit a public comment for the first time. Not because she thought one comment would solve everything, but because it made her realize that her expertise had value in those conversations. And she is right.

We have the expertise to speak to the methodological consequences of eliminating surveys. We understand the downstream effects of losing data. We know what disappears when documentation is removed, when staffing is cut, when preservation workflows break down, or when access becomes restricted.

And we have a professional obligation, I would argue, to say so out loud, in public, in the places where those decisions are being made. Not because we are abandoning our professional roles, but because we are fully embodying them. When we talk to administrators, legislators, journalists, and to the public, we need to be willing to say clearly: this is not just about research. This is about democracy. This is about the public’s right to know. This argument runs through this entire conference. I encourage you to look at the cluster of Indigenous data sovereignty sessions in Session F.1. These sessions are insisting, clearly and forcefully, that data access is a rights issue.

Especially in an era increasingly shaped by AI-generated information, our profession’s long-standing attention to provenance, context, documentation, transparency, and access becomes even more important. These are not peripheral concerns. They are central to how people determine what information they can trust. We should not underestimate the importance of having a visible professional voice in those conversations. The Data Rescue Project showed me that people are already looking to this community for leadership, whether we recognize it or not.

IASSIST Sessions

Many of this week’s sessions reflect this direction. I want to flag just a few.

Cameron Blevins and colleagues will present the Data Advocacy for All Toolkit in Session A.1, which is exactly the kind of resource that can help us act on this expanded sense of purpose. Annelise Sklar will talk about maintaining a licensed data collection amid unexpected budget reductions in Session C.2 — a situation I suspect many of you’ve faced or are facing. And Mikala Narlock — one of my DRP Steering Committee colleagues — will present on Championing Sustainable Data Services: Strategic Scoping as Advocacy in Session J.1. Tomorrow I’ll chair a panel titled Managing Through Crisis: Supervising Data Librarians in Uncertain Times, where we’ll talk frankly about what it means to lead data services departments in this environment. These sessions, and others, are not departures from what we do. They are broader expressions of it.

Returning to our IASSIST Community

When I started as a data librarian in 2007, I had three local mentors who pushed me to become part of this community: Michele Hayslett, Joel Herndon, and Jim Ovitt. Jim passed away just a year later, at 51. All three of them told me to go to IASSIST. And I was nervous, because my previous academic conference experiences had not been particularly welcoming. As a PhD student in political science, I found that conferences tended to be more about competition than community. But IASSIST was different. Even at my first conference, which was heavy on DDI and metadata, much of which I did not fully understand, I was struck by how genuinely welcoming people were. How willing they were to share what they knew. How much they seemed to actually like each other.

That community has been a constant throughout my career. It has shifted as people move on or change roles, but it has never gone away. And when 2025 arrived with everything it brought, that community showed up.

The crises we face are real. The funding crisis is real. The capacity problem is real. The political and technological disruptions are real. I am not going to stand here and tell you it is not hard, because it is.

But someone in this virtual room right now is at their first IASSIST. Someone who does not know everyone yet and who is not sure they belong here. There is someone who will, over the next decade or two, become one of the people who hold this community together. And we need to think about what we pass to that person. Not just our technical knowledge, but our understanding of why this work matters, our belief that data professionals have a civic role to play, and our conviction that access to information is essential to democracy. This is the work that outlasts any single crisis. This community has been building something real for decades, person to person, conference to conference. And I am grateful to be building it with you this week. That is the reason for hope in dark times.

Thank you. I look forward to maybe seeing your cats and dogs on Zoom screens this week. And I wish you all the best IASSIST ever.

–

Slides are available on Zenodo: Kellam, L. (2026, June 4). Championing Public Data: Data Professionals in Civic Life. IASSIST 2026 Virtual Conference. Zenodo. https://doi.org/10.5281/zenodo.20547211