Exploring the IASSIST 2017 program: a guide to key sessions

The Lawrence campus of the University of Kansas played host in late May 2017 to one of the year's most anticipated gatherings of research data professionals. Under the banner "Data in the Middle: The Common Language of Research," IASSIST 2017 brought together librarians, data curators, statisticians, and domain scientists for four days of intensive programming. Sessions ran from May 23 through May 26, with pre-conference workshops stretching the conversation back into the preceding days. The theme was deliberately broad, designed to bridge the methodological divides that often separate qualitative researchers from quantitative specialists, archivists from active data producers.

IASSIST itself, the International Association for Social Science Information Service and Technology, has long occupied a niche at the intersection of data stewardship and applied research. Founded in the 1970s, the association had by 2017 steadily broadened its scope beyond its original social science base. The program reflected that evolution: agronomists, computer scientists, and policy analysts shared the same rooms as data librarians. The Lawrence venue, set along the Kansas River and within easy reach of the historic downtown, offered a walkable campus where breakout sessions sat a stone's throw from the main plenary hall.

For an Australian audience, the conference carried particular resonance. Researchers in Brisbane, Melbourne, and Canberra have spent the past decade building national data infrastructure through the Australian Research Data Commons and its predecessor, the Australian National Data Service. Themes of data citation, persistent identifiers, and metadata interoperability, staples of any IASSIST meeting, map directly onto conversations happening at the University of Sydney, the Australian National University, and the University of Melbourne. The 2017 program offered a useful snapshot of where international thinking stood at a pivotal moment for data-intensive scholarship.

The remainder of this guide walks through the structure of the meeting, from the opening plenaries through to the closing workshops, and highlights the kinds of sessions that tend to leave the longest impression. It also touches on the practical side of attending, with a few notes drawn from the experience of registrants who travelled from as far afield as Adelaide and Perth. Whether you are revisiting the archived program or discovering it for the first time, the sections below should help you find sessions that suit your own research interests.

The conference theme and programmatic spine

"Data in the Middle" was more than a slogan; it signalled an editorial commitment to foreground the connective tissue between data creators, intermediaries, and users. The program committee, chaired by figures from the University of Kansas Libraries, structured the week around several thematic strands. Plenary talks anchored the mornings, while parallel sessions in the afternoons allowed attendees to mix and match across data domains. Lightning rounds, poster sessions, and Birds-of-a-Feather gatherings filled the late afternoons.

The schedule reflected the realities of organising a conference for researchers juggling competing demands. A first-timer from the Melbourne data archive team once noted that the spread of topics at IASSIST makes it possible to learn from a librarian one hour and a deep-learning engineer the next. That breadth is part of the association's long-running appeal. Below is a simplified overview of how the four main days were organised, drawn from the archived program.

Day	Main focus	Featured topic areas
Tuesday, May 23	Opening day and plenary	"Data in the Middle" keynote, welcome reception
Wednesday, May 24	Big data and computational methods	Deep learning, text mining, scalable analytics
Thursday, May 25	Domain applications	Digital agriculture, global food security, climate datasets
Friday, May 26	Closing day and synthesis	Future directions, community reports, closing remarks

The table condenses a packed program into its core rhythms. Wednesday and Thursday carried the heaviest concentration of domain sessions, while the bookend days focused on framing, plenaries, and synthesis. Attendees planning ahead generally prioritised those midweek days for their must-see talks, leaving Tuesday and Friday for panel discussions and informal catch-ups.

Plenary sessions and keynote highlights

Each morning of the main conference opened with a plenary designed to set a shared intellectual agenda. The 2017 keynote speakers were chosen to embody the conference's cross-domain ambitions. One drew on work in deep learning applied to large-scale text corpora, illustrating how neural network approaches were beginning to reshape research in the social sciences. Another addressed the politics of open data in the context of agricultural research, drawing examples from collaborative projects spanning the Americas, Africa, and Asia.

The plenary format at IASSIST has historically favoured speakers who can speak to both methodological detail and institutional strategy. In the Australian context, where bodies such as the Australian Research Council have pushed for FAIR data principles, the strategic questions raised in these plenaries feel immediately relevant. Researchers attending from the Group of Eight universities often find that the keynote discussions echo conversations at their home institutions.

Beyond the formal keynotes, the conference featured a panel on the future of data citation, an issue that has consumed considerable energy at Australian institutions. The conversation touched on persistent identifiers, the work of the Research Data Alliance, and the practical difficulties of getting busy researchers to mint DOIs for every dataset. These plenary conversations frequently set the tone for the technical sessions that followed in the afternoon slots.

Big data and deep learning tracks

Wednesday's program leaned heavily into computational methods, with multiple streams dedicated to big data analytics and machine learning. The deep learning strand attracted a strong turnout, reflecting a broader trend across the research data community. Researchers presented work on convolutional networks for image classification, recurrent architectures for sequence modelling, and the practical challenges of training such models on sensitive social science data.

For attendees from institutions such as the University of Queensland or the Australian National University, several of these sessions offered benchmarks against which their own internal initiatives could be measured. Presentations on differential privacy, federated learning, and ethical review of algorithmic pipelines resonated with work being done under the auspices of the Australian Academy of Science. A common thread was the gap between the capabilities demonstrated in well-funded computer science departments and the day-to-day reality of researchers in smaller teams.

The big data track also included a session on working with administrative data, an area in which Australia has built significant expertise through initiatives like the Multi-Agency Data Integration Project. Presenters from both sides of the Pacific shared practical lessons on linking records across agencies while preserving confidentiality. The format mixed short presentations with structured discussion, allowing participants to draw out differences in regulatory environments that often shape what is technically possible.

Digital agriculture and global food security

Perhaps the most distinctive thematic strand of IASSIST 2017 was the focus on digital agriculture and global food security. Held on Thursday, this stream attracted researchers from agronomy, remote sensing, food policy, and rural sociology. Sessions explored how satellite imagery, sensor networks, and genomic data were reshaping the way scholars approach questions of crop yield, food distribution, and rural livelihoods.

The choice to foreground agricultural research was particularly timely. In 2017, conversations about climate adaptation, sustainable intensification, and smallholder resilience were gaining prominence across international research funding bodies. For Australian researchers accustomed to working on dryland agriculture in regions such as the Murray–Darling Basin or the southern rangelands of Western Australia, the sessions on water-efficient cropping systems and remote sensing of soil moisture felt immediately applicable.

Several presentations examined how open data platforms were being used to coordinate research across borders. One traced the work of the CGIAR consortium in standardising agricultural metadata, while another looked at how US university extension services were integrating public datasets into outreach. The Australian angle surfaced in discussions of the Atlas of Living Australia and similar biodiversity informatics efforts. By bringing these threads together, the conference made a case for treating food systems data as a serious object of cross-disciplinary inquiry.

Workshops, tutorials, and hands-on learning

Outside the main plenary and themed sessions, the program included a substantial workshop component. Pre-conference tutorials on May 22 covered topics such as data wrangling in R, introductory Python for data analysis, and metadata standards including Dublin Core, DDI, and DataCite. These workshops were ticketed separately and typically filled within weeks of registration opening, reflecting strong demand from early-career researchers and data librarians looking to refresh their technical toolkit.

The hands-on format suited the IASSIST audience. Many attendees arrived with specific problems from their own institutional contexts and used the workshops to test new tools. Participants from Australian universities frequently report that the metadata workshops are where they pick up techniques they later bring back to local research data teams. The collegial atmosphere of the tutorial rooms, with small groups and patient facilitators, tends to produce lasting professional connections.

Mid-conference workshops were also embedded into the main program. These typically ran for three hours and covered topics such as data visualisation with Tableau, geospatial analysis with QGIS, and reproducible research workflows using Jupyter notebooks. The conference organisers were careful to ensure that workshops did not simply replicate content available in online courses, emphasising instead the curated expertise of facilitators and the opportunity to ask detailed questions.

Practical logistics and recommendations

The practical side of attending deserves attention. The University of Kansas provided a venue map, accommodation blocks, and detailed transport guidance through the conference website. Registration opened several months in advance, with discounted rates for early-career researchers and students, a model now common across professional associations in the data space.

Attendees flying in from outside the United States typically routed through Kansas City International Airport, with ground transport options to Lawrence documented on the conference site. Those arriving by rail could use the Lawrence Amtrak stop, a short walk from the campus accommodation. Within Lawrence itself, the conference footprint was compact enough that most attendees did not need a car, an arrangement familiar to anyone who has spent time in the inner suburbs of Sydney or Melbourne.

The downtown Lawrence area offers an unexpected density of independent bookshops, coffee roasters, and live music venues, which provided a relaxed social backdrop for evening networking. The official registration page remains a useful resource for understanding the fee structure, the workshop pricing, and the kinds of support available for attendees travelling on restricted research budgets.

Readers exploring the archived program with a view to future professional development may find the following suggestions useful. They are drawn from the kinds of decisions that seasoned attendees tend to make when planning their week at a large multi-track conference.

• Build a personal schedule that mixes two or three thematic strands rather than chasing a single track; the cross-pollination at IASSIST is part of its value.
• Reserve at least one evening for the poster sessions and Birds-of-a-Feather meetings, where conversations often lead to longer-term collaborations.
• Treat the pre-conference workshops as an investment rather than an add-on, particularly if you are relatively new to data management tools.
• Read the plenary abstracts in advance and prepare one or two questions, since the Q&A sessions tend to be richer when attendees arrive with focused points.
• Reach out to fellow attendees from your region before the conference, because Australian researchers often coordinate travel and shared accommodation to manage costs.
• Capture notes during sessions and share them afterwards through institutional channels, as many of the most valuable insights come from informal hallway discussions.
• Allow time after the closing session to decompress, reflect, and follow up with new contacts; the day after a conference is often the most productive part of the trip.