A practical guide to the IASSIST 2017 Data Fair

Between 23 and 26 May 2017, the University of Kansas in Lawrence hosted “Data in the Middle: The Common Language of Research,” the IASSIST conference. The Data Fair gave the event a hands-on dimension through research datasets, visualisations, software demonstrations and conversations about how evidence moves between disciplines.

The archived programme is now a historical guide rather than a current visitor itinerary. It remains useful for understanding how IASSIST connected researchers, practitioners and domain specialists, while also showing how ideas about big data, reproducible research, digital agriculture and responsible data use can be applied in Australia.

The Data Fair as a shared research space

IASSIST brings together statisticians, data scientists, researchers and practitioners working with quantitative information. Its 2017 theme was especially relevant to a Data Fair because research data can become inaccessible when technical language separates specialists from the people who use their findings. Demonstrations, posters and presentations created pathways for explaining evidence across disciplinary boundaries.

The fair brought major programme themes into direct conversation. Big data systems, deep learning methods, digital agriculture and global food security each addressed a different kind of scale, from computational volume to environmental and social impact. Visitors could compare analytical approaches while examining the metadata, provenance and communication choices that made results understandable.

For an archived event, the programme works as a thematic guide rather than a live exhibitor directory. Its strongest contribution was showing how a fair can connect a dataset with its source, an algorithm with its assumptions, and a research finding with a real-world decision. That connection remains the defining feature worth preserving when recreating or evaluating a data-focused conference.

Finding your way around Lawrence

The archived practical information covers check-in, transportation and recommendations for the local area. Lawrence has a compact university-town setting, with campus buildings, hospitality venues and downtown businesses connected by walkable routes and local transport. Visitors using an archived schedule should confirm the precise Data Fair location and check-in arrangements before relying on a map or historical timetable.

Australian city travellers will recognise familiar transport habits. Sydney passengers tap Opal cards for trains and buses, Melbourne travellers use Myki, and Brisbane journeys rely on TransLink. In Lawrence, the equivalent routine was generally simpler, with walking and short rides suiting the compact campus. The comparison matters because Australian conference organisers accustomed to metropolitan travel should still provide clear campus maps, accessible routes and realistic transfer times.

Clock differences also shape a conference day. During May, Sydney and Brisbane were about 15 hours ahead of Lawrence, while Adelaide and Darwin were about 14½ hours ahead. Remote participants and international attendees needed to translate plenary times carefully, especially when arranging demonstrations or meetings with presenters. Keeping a converted schedule, a power source and contact details for presenters reduced unnecessary friction.

Reading the programme as a research map

The conference schedule can be read in layers: plenary sessions supplied broad context, parallel sessions developed specialist ideas, and the Data Fair showed those ideas in practice. Plenary material is particularly useful for visitors because it explains the common vocabulary running through the event. After a keynote, a participant could recognise recurring concerns about data quality, statistical inference, visualisation, knowledge sharing and reproducibility.

The programme’s subject areas also form a connected research map. Big data sessions concerned heterogeneous, continuously generated information; deep learning sessions addressed model design, training data and computational interpretation. Digital agriculture combined sensing, prediction and decision support, while global food security linked those tools to production, distribution, access and policy. Shared concepts such as metadata, validation, uncertainty and reproducibility connected these apparently different domains.

A productive approach was to alternate between broad and specialist sessions. Start with a plenary, record the concepts that recur, then choose domain sessions that demonstrate those concepts in use. Allow time around the Data Fair for comparison: one demonstration might emphasise a novel algorithm, another a public-service application, and another the communication of uncertain results. Questions at the end of a talk often became more productive after seeing the underlying evidence in a demonstration.

Learning from the data demonstrations

A strong Data Fair demonstration reveals more than a polished interface. Look for the source data, preprocessing steps, analytical assumptions and intended audience. Ask how missing values were handled, whether the example could be reproduced, and which conclusions remain uncertain. These questions turn a visual display into a research narrative and help visitors assess practical value without confusing technical sophistication with reliable evidence.

Digital agriculture provided an especially clear example of common research language. Satellite imagery, soil measurements, weather records, yield estimates and supply-chain information can inform decisions about planting, irrigation, harvesting and food distribution. The global food security theme connected those technologies to questions of affordability, resilience and access, ensuring that faster or more precise analysis did not become an isolated technical achievement.

Deep-learning demonstrations similarly required context around training data, labelled examples, model validation and bias. The atmosphere at a 2017 fair understandably centred on capability, yet current interpretation should also consider privacy, explainability, reproducibility and unintended effects. The archived material shows how quickly public expectations can change, making historical conference records valuable evidence for developing better standards for analytical demonstrations.

Translating the event for Australia

Australia’s consumer market offers a useful parallel. Coles, Woolworths and Aldi operate extensive loyalty programmes, while agricultural businesses generate data through precision farming, machinery, weather services and supply-chain systems. These settings raise familiar questions about consent, commercial use, data ownership and responsible sharing. The experience of explaining research at the Data Fair helps researchers communicate those questions when working with industry, government or community partners.

Everyday digital habits make data stewardship tangible. Australians commonly use mobile banking, supermarket applications, public transport cards, digital health services and online accounts that leave traces across multiple organisations. These habits show why data literacy must extend beyond technical vocabulary. People need practical information about who receives their information, how it is protected and how they can exercise their rights under Australian law.

Privacy law provides a firm framework for contemporary adaptations. The Privacy Act 1988 and the Australian Privacy Principles govern many uses of personal information, while the Office of the Australian Information Commissioner offers guidance for organisations. The Spam Act 2003 also affects electronic marketing practices. Australia’s Notifiable Data Breaches scheme began in February 2018, after the conference, so a modern event should distinguish the legal position in 2017 from present requirements. The archived conference lessons provide a valuable event-level companion for evaluating how practical context, attendee support and shared learning shaped the original gathering.

Research touching Aboriginal and Torres Strait Islander communities requires additional care. Present-day Australian data fairs should consider the AIATSIS Code of Ethics, Indigenous data governance and community authority over information. A technically open dataset can still cause harm if culturally sensitive information is published without appropriate consent, consultation or control. Demonstrations should therefore include provenance and stewardship alongside computational performance.

A reusable model for future data fairs

The most durable lesson from the 2017 Data Fair is that interoperability requires communication, not merely compatible file formats. A useful event gives specialists and general audiences enough shared language to discuss evidence without erasing the context of each discipline. It also gives researchers a place to show limitations, workflows and potential applications rather than presenting finished claims without explanation.

A compact comparison connects the archived themes with present-day Australian practice:

Archived emphasis	Useful question for visitors	Australian application
Big data	Where does the information originate, and how is it cleaned?	Open government data, transport systems, consumer research and environmental monitoring
Deep learning	Which training data, labels and evaluation methods shaped the result?	Reproducible machine learning with privacy, bias and explainability checks
Digital agriculture	How does sensing improve an operational decision?	Drought monitoring, irrigation, crop forecasting and resilient supply chains
Global food security	Who benefits from better analysis, and who might be excluded?	Food access, affordability, regional resilience and community-led research
Shared research language	Which definitions and metadata allow methods to be compared?	Consistent stewardship across universities, agencies and industry
Plenary and parallel sessions	How do broad ideas connect with specialist evidence?	Coordinated hybrid programmes with shared questions and accessible online access

Read each row as a prompt for examining the fair’s design, scientific content and relevance. The strongest exhibits would have answered the visitor questions while revealing their limits. A modern version should preserve that transparency, allowing attendees to understand both what research can do and what responsible use requires.

For anyone preparing a comparable visit, the archived material supports a focused preparation routine:

• Browse the conference schedule and locate plenaries, parallel sessions and Data Fair activities.
• Select at least one methodological session and one domain-focused session for each major theme.
• Review the source, consent and governance details of any dataset used in a demonstration.
• Convert the relevant Australian time zone into notes, reminders and meeting requests.
• Use the practical pages to understand historical check-in, transport and local-area guidance.
• Prepare specific questions about reproducibility, uncertainty, limitations and intended users.

During the fair, a structured routine helps balance technical depth with wider discussion:

• Begin with a plenary to establish the shared vocabulary used across the programme.
• Compare demonstrations by data source, analytical method, audience and stated limitations.
• Ask each project how its results could be validated, reproduced or independently audited.
• Connect digital agriculture and food-security work with Australian environmental conditions.
• Record project names, collaborators and follow-up resources for relevant research.
• Leave time for informal conversations that may reveal applications absent from the formal sessions.

Used this way, the 2017 Data Fair is more than a historical programme; it is a working model for turning complex research into public value.