Bridging the digital divide in agriculture
The archived IASSIST 2017 conference, “Data in the Middle: The Common Language of Research,” brought researchers and practitioners to the University of Kansas in Lawrence from 23 to 26 May 2017. Across its programme, plenaries and presentations examined big data, deep learning, digital agriculture and global food security, showing how shared data languages can connect laboratory work with field decisions.
Digital inequality in farming extends beyond a lack of internet service. It also appears in expensive sensors, limited computing capacity, weak data-management skills, restrictive licences and platforms designed around large commercial operations. Smaller farms and remote communities may contribute local observations that are indispensable to research yet receive little analytical capacity or decision-making power in return.
For an Australian audience, the archive remains useful because it captures a moment when agricultural data moved from specialist projects toward everyday farm management. Its themes translate into practical questions about coverage, interoperability, ownership and participation.
Who experiences the agricultural digital divide
The conference presentations framed the divide as a chain of access rather than a single barrier. A farmer may be able to download a weather application but lack the connectivity to upload detailed sensor records continuously. A researcher may have a server for machine-learning experiments while a grower lacks a usable dashboard or the time to maintain it. Each gap can prevent evidence from becoming a better forecast, irrigation plan or harvest decision.
Language and representation deepen the problem. Data collected by large, well-funded projects can become the default evidence base, while smallholder conditions, seasonal variability and local ecological knowledge remain poorly described. Global food security discussions showed why this matters: models trained on narrow evidence may recommend practices that do not transfer to another soil, climate, crop or production system.
Trust also shapes participation. Farmers are more likely to contribute records when they understand who will use them, whether they can reuse them and how benefits will be shared. Open data alone does not answer those concerns. Consent, clear governance and relevant feedback are needed if data infrastructure is to serve both research and the people whose work generated the evidence.
Making field data usable for longer
Long-term field experiments depend on documentation that survives staff changes, grant cycles and equipment upgrades. The archived material on field data management highlights the practical value of consistent metadata, managed vocabularies and durable repositories. Without those elements, years of careful measurements can become isolated files that are difficult to interpret even by the original team.
Digital agriculture can generate extraordinary volumes of observations from soil probes, weather stations, satellite imagery and machinery. Scale creates value only when records can be linked, checked and compared. Naming conventions, variable definitions and quality flags allow a researcher to distinguish a sensor outage from a real change in soil moisture, avoiding a false signal that might influence a recommendation.
The conference’s emphasis on a common research language supported a practical principle: preserve context alongside measurements. Location, crop variety, management history, calibration and collection method can matter as much as the number itself. That principle is essential when Australian trials move between the Darling Downs, the northern Victorian cropping belt and pastoral areas where rainfall patterns differ sharply.
What deep learning can and cannot reveal
Deep learning received prominent attention as a method for image recognition, prediction and pattern detection. For agriculture, these tools can help identify crop conditions from drone photographs, classify produce or combine sensor streams into an early-warning signal. Such systems can reduce the time required to inspect large areas and make underused data more informative.
Training data sets the boundary of those benefits. If a model has seen mainly large, mechanised farms in particular regions, it may perform poorly on small plots, unusual cultivars or mixed cropping systems. The conference’s global food security framing encouraged attention to missing evidence, because scale can conceal bias rather than remove it. More observations are helpful when they represent the environments in which decisions will be made.
Interpretability and cost also affect adoption. A grower needs to know why a recommendation appeared and what uncertainty surrounds it. Cloud processing may be economical for a large operation but unsuitable where bandwidth is limited or a subscription becomes unaffordable. Local, low-power or edge-based tools may offer a better fit, provided their accuracy and maintenance needs are tested in real conditions.
Standards, ownership and shared infrastructure
Interoperability emerged as a bridge between isolated projects and wider collaboration. Shared standards can let a soil reading from one trial be compared with a weather record from another, while common identifiers can reduce duplication. This matters in publicly funded research, where agencies, universities and private agronomy firms may all hold relevant records but use incompatible systems.
Ownership must accompany interoperability. A technically open feed can still be enclosed by terms that restrict reuse, or by a platform that holds data in a form customers cannot export. Conference discussions pointed toward governance as part of infrastructure: agreements should cover access, attribution, privacy, security, retention and the return of useful results to participants.
Shared infrastructure can be more effective than a single proprietary service. Universities can host repositories, government agencies can provide reference datasets and standards, while farmer organisations and technology suppliers test interfaces in daily use. This layered approach can serve a research community while leaving room for commercial innovation and local adaptation.
Food security meets Australian farm realities
Global food security depends on information as much as water, seed or transport, yet information can follow an unequal path from field to policy. The IASSIST discussions showed that research capacity, infrastructure and participation must be considered together. A dataset that cannot travel, or a forecast that cannot be interpreted locally, does not close the gap between productive regions and those facing more persistent constraints.
Australian conditions make that lesson immediately practical. Around Moree and Dubbo in regional New South Wales, as well as Toowoomba in southern Queensland, producers may have usable mobile coverage while fixed fibre ends at the township boundary. The limiting factor may be power, equipment, technical support or a service subscription rather than the location of mobile towers. One rural property can therefore be more connected than the next within the same district.
Trade and direct selling add further layers. Australian customs and biosecurity requirements affect the movement of seed, livestock, machinery and produce, making dependable traceability important across public agencies and private supply chains. Producers selling through Melbourne’s Queen Victoria Market, Adelaide Central Market or Sydney Fish Market may need clear provenance, yet small vendors can lack the digital systems available to large wholesalers. Portable records and cooperative services can help without forcing every stallholder to buy complex software.
A local perspective also matters in northern and remote Australia. Partnerships involving First Nations communities should follow appropriate protocols for Indigenous data governance, consent and benefit sharing. Digital tools should strengthen community priorities rather than extract information from country for outside use. Resources such as Unleash Opportunity offer related perspectives on connecting digital inclusion with community capability and opportunity.
Turning conference lessons into practical action
The conference archive supports a layered response rather than a single technology purchase. It treats access as something to be designed with farmers, researchers, advisers and public agencies, while recognising that every group enters the process with different resources and obligations.
Australian conditions to account for include:
- Connectivity differs sharply between metropolitan services and farms beyond towns such as Melbourne, Brisbane, Adelaide and Perth, making field connectivity and offline access essential design requirements.
- Australian farms range from large grain and livestock enterprises to small family operations, so a tool priced only for commercial scale can deepen the divide.
- Customs and biosecurity workflows create demand for traceable records, including situations where paper documentation and digital exchange must coexist.
- Melbourne’s Queen Victoria Market, Adelaide Central Market and Sydney Fish Market offer direct-market opportunities, but provenance tools must remain affordable for small stalls and producers.
- Remote and First Nations communities require data governance shaped by local authority, language requirements and accountable benefit sharing.
The conference legacy translates into five practical actions:
- Begin with a local decision or research question, then identify the smallest reliable dataset needed to support it.
- Combine better connectivity with equipment access, training, technical support and affordable ongoing services.
- Use open standards, durable metadata and clear data agreements so evidence can be combined without erasing context.
- Design low-bandwidth and mobile-friendly services, with human advice available when automated interpretation is uncertain.
- Measure participation and benefit across farm sizes, locations and demographic groups, then adjust programmes using that evidence.
These actions connect the conference’s research emphasis to service design. They also acknowledge that digital inclusion is sustained work: repositories need maintenance, data agreements need stewardship and users need confidence that their contributions produce visible value.
Comparing inclusive models for agricultural data
No single model resolves the agricultural digital divide. Connectivity programmes address physical access but cannot by themselves ensure that records are trusted or useful. Open-data policies improve reuse, yet they can overwhelm small producers with consent forms or technical requirements. Commercial services often provide polished interfaces, but pricing and platform control may exclude smaller users. Community-led and public shared approaches can fill these gaps when funding and governance are durable.
The conference archive is most valuable as a prompt for making those choices collectively. Plenary themes, technical sessions and discussions on global food security reveal a common need: research infrastructure should be legible to the people who contribute observations and responsive to the decisions they must make. An Australian programme can apply that lesson by combining national standards with regional delivery, formal biosecurity systems with accessible market tools, and advanced analytics with straightforward advice.
| Model | Main strength | Common limitation | Australian application |
|---|---|---|---|
| Connectivity expansion | Enables devices, maps and remote support | Does not guarantee skills, affordability or useful data | Pair fibre and mobile improvements with equipment and advisory services |
| Open-data cooperative | Shares evidence and reduces duplication | Requires governance, metadata work and participant trust | Build farmer and research cooperatives around common standards |
| Skills and co-design | Produces tools suited to local decisions | Takes time and sustained facilitation | Design with regional advisers, small farms and community organisations |
| Commercial platform | Offers convenient interfaces and specialist support | Can create subscription barriers and lock-in | Require exportable data, transparent terms and accessible support tiers |
| Shared public infrastructure | Connects research, agencies and industry | Needs stable funding and long-term stewardship | Combine repositories, reference datasets and extension services |
At the Conference
What attendees experienced in Lawrence
Plenary Sessions
Keynotes from Daniel Reed on data, technology, and culture, and Jennifer Clarke on digital agriculture and the Midwest Big Data Hub.
Workshops
Full-day technical workshops on Tuesday, May 23. Morning sessions ran 9:00–12:00 and afternoon sessions 13:00–16:00. Laptops were required.
Social Events
An opening reception at The Oread, a banquet, and an optional post-conference tour of Kansas City including Crescent Moon Winery.
Venue & Accommodations
Where the conference took place
Kansas Union
University of Kansas campus, Lawrence. Main conference venue with check-in on the 4th and 5th floor lobbies.
The Oread
1200 Oread Avenue, Lawrence. Hosted the opening reception and offered a room block for attendees.
The Eldridge
701 Massachusetts Street, Lawrence. A partner hotel with a reserved room block for conference guests.
Springhill & TownePlace Suites
Marriott properties in Lawrence with room blocks reserved under the "KU IASSIST Conference" name.
Program Highlights
Sessions and activities
Getting Here
Lawrence, Kansas
Kansas Union · University of Kansas
1301 Jayhawk Blvd, Lawrence, KS 66045
Kansas City International Airport (MCI) is approximately 50 minutes by car. Lawrence Transit Routes 10 and 11 served the area ($1 exact change).
Plan Your Stay
Accommodation options that were available
The Eldridge
701 Massachusetts Street, Lawrence, KS 66044. Room block now closed.
The Oread
1200 Oread Avenue, Lawrence, KS 66044. Room block now closed.
Springhill Suites
Marriott property. Room block reserved under "KU IASSIST Conference."
TownePlace Suites
Marriott property. Room block reserved under "KU IASSIST Conference."