Post-Conference Data Archiving For Research Presenters

A conference presentation usually ends when the slides are delivered, questions are answered and the audience leaves the room. The research itself has a longer life. Datasets, code, interview instruments, figures and explanatory notes may continue to support replication, teaching, policy work or a later publication. Without deliberate preservation, however, valuable material can become difficult to interpret within a few years.

Post-conference data archiving gives presenters a practical way to turn a temporary research output into a durable scholarly resource. The archived IASSIST 2017 programme, held at the University of Kansas in Lawrence from 23–26 May 2017, is a useful model: a conference record becomes more valuable when its sessions, presentations and supporting information remain discoverable after the event has finished.

Why The Archive Matters

A presentation deck is rarely a complete account of a research project. It may contain shortened labels, unexplained acronyms, selected results and visualisations designed for a live audience. An external researcher who encounters the file later may not know which dataset produced a chart, what version of the code was used or whether the findings were preliminary.

A well-prepared archive preserves the relationship between the presentation and its evidence. It can include the final slides, an abstract, a dataset description, analytical scripts, a codebook, survey instruments, consent information, outputs and a README file. These materials give future users enough context to understand what was done without treating an old presentation as if it were a complete methods paper.

Archiving also protects institutional memory. Australian research teams often move between universities, government departments, consultancies and industry partners. Staff may change roles, laptops may be replaced and cloud accounts may be closed. A carefully documented deposit reduces the risk that a project depends on one researcher’s personal hard drive or an inaccessible team folder.

The goal is not to publish every file. It is to preserve the right files, explain their status and record any limits on reuse. A restricted dataset, a de-identified extract or a synthetic example can still support transparency when the original material cannot be openly released.

Audit And Prepare The Research Package

Begin with an inventory of everything connected to the presentation. Include the submitted abstract, speaker notes, final slides, posters, figures, tables, raw and processed data, scripts, software environments, analysis logs and related publications. Record the current location, owner, format, date and sensitivity of each item. This simple audit often reveals duplicate files with confusing names such as final_v3_revised_reallyfinal.xlsx.

Separate working material from the authoritative version. Keep drafts when they document an important analytical decision, but identify which file should be cited. Use consistent filenames that communicate the project, content, version and date, such as river-quality-code-v1.2-2024-08-16.R. Dates should follow an unambiguous format, preferably YYYY-MM-DD, because Australian dates written as day-month-year can be misread by international collaborators.

Check whether the presentation has changed since it was delivered. A later correction, reprocessed dataset or revised figure should be recorded in a changelog rather than silently replacing the original. If a result shown at IASSIST 2017 was exploratory, label it as such. Preservation should maintain an honest record of what was known and presented at the time.

Before deposit, remove temporary files, passwords, hidden spreadsheet tabs, tracked changes and embedded personal information. Inspect PDFs and slide decks for speaker notes or comments that were not intended for public release. Export proprietary formats into stable alternatives where practical, while retaining the original working file when it is needed for reproducibility.

Apply Australian Governance And Ethics

Data stewardship in Australia must account for legal, contractual and community obligations. The Privacy Act 1988 and the Australian Privacy Principles may apply to personal information, although coverage depends on the organisation and circumstances. Universities and health services may also impose stricter policies. A dataset that seems harmless after removing names can still identify people through location, occupation, dates, rare conditions or combinations of variables.

Review the original consent wording before sharing research data. Participants may have agreed to use within a specific project, institution or time period without agreeing to open publication. For health, education, employment and social research, the archive may need to contain only a metadata record, a controlled-access request process or a carefully de-identified subset. Document who can approve access and what evidence an applicant must provide.

Indigenous data requires particular care. Australian researchers should consider Indigenous Data Sovereignty and Indigenous Data Governance principles, including community authority over collection, interpretation, access and reuse. A conventional open-data licence may be inappropriate if it ignores cultural rights or community expectations. Where Aboriginal or Torres Strait Islander knowledge, places or communities are represented, governance arrangements should be recorded alongside technical metadata.

Copyright can affect slides, photographs, maps, survey instruments, software and third-party datasets. The Copyright Act 1968 does not automatically give a presenter permission to redistribute material merely because it appeared in a conference talk. Confirm whether images, commercial data or co-authored outputs can be deposited, and use a licence that reflects the permissions actually available.

Researchers working on agriculture, drought, food supply or regional development should also consider whether precise coordinates or operational details could create risks. A farm location, small-community result or infrastructure map may need geographic aggregation. For context on how food-security research was discussed in the conference archive, presenters can review these food security sessions before deciding which supporting materials belong in a public record.

Build Metadata That Explains The Evidence

Metadata is the information that makes archived research intelligible. At minimum, describe the title, creators, affiliations, abstract, keywords, date, funder, methodology, geographical area, temporal coverage, data sources, file formats, access conditions and related publications. Add a persistent identifier such as a DOI when the repository provides one, and cite the archive in future papers and presentations.

Write for someone who was not in the room. Explain what each dataset contains, how variables are coded, which values represent missing data and whether units are metric. A README should identify the software required to open or analyse each file, the order in which scripts should be run and the expected outputs. If an analysis depends on a particular package version, operating system or external API, record that dependency.

Use controlled vocabularies and discipline standards when they improve discovery. Keywords such as “digital agriculture”, “deep learning”, “food security”, “research data management” and “spatial analysis” can connect a deposit with related work, but they should be supported by precise terms describing the actual material. Avoid filling metadata with broad promotional language that does not help another researcher assess relevance.

Document uncertainty and limitations. State whether values were imputed, whether the sample was representative, how missing observations were handled and whether the presentation contains preliminary findings. If a source dataset cannot be shared, explain its provenance and provide enough information for an authorised researcher to request access or obtain an equivalent source.

Good metadata should also identify relationships. Link the presentation to the dataset, software, article, grant, conference session and later correction. This creates a research object rather than a disconnected collection of files. A repository record that clearly connects these elements is easier for libraries, search engines and future collaborators to interpret.

Select A Repository And Deposit Carefully

Choose a repository that matches the material and its access requirements. An institutional repository may be best for slides and conference papers, while a disciplinary data service may provide stronger support for geospatial, social science, health or agricultural datasets. Government-funded projects may have funder-specific rules, and some universities require deposit in an approved service.

Assess the repository’s preservation practices before uploading. Look for persistent identifiers, version control, access controls, metadata export, file-size limits, retention commitments and clear terms of use. A platform that offers only a download link may be convenient but weak for long-term stewardship. Check whether the repository can preserve checksums, maintain fixity and notify you when a file is replaced or withdrawn.

Prefer open, documented formats where they preserve meaning. CSV works well for simple tabular data, though it needs a data dictionary and an agreed character encoding. TXT, XML, JSON, TIFF, WAV and PDF/A may be suitable in different contexts. Proprietary formats such as SPSS, Stata, ArcGIS or specialist laboratory files can remain useful, but provide an export or explanatory alternative where possible. Do not convert complex data blindly: an export that loses labels, formulas, coordinate systems or relationships is not a preservation success.

Choose access conditions deliberately. “Open”, “embargoed”, “restricted” and “metadata only” are different decisions. An embargo should have an end date and an explanation. Restricted data should identify the custodian and access route. If a dataset has no lawful or ethical pathway for release, its catalogue record can still improve discoverability without exposing the content.

After deposit, test the record as an outsider. Download the files, open them on a separate device and follow the README without relying on local knowledge. Check that links work, the citation resolves, the licence is visible and the files match the description. Ask a colleague who was not involved in the project to identify any ambiguity.

Archiving Need	Suitable Material	Useful Preservation Choice	Access Consideration
Explain the presentation	Slides, abstract, poster, speaker notes	PDF/A plus an editable source file where permitted	Open, unless notes contain sensitive content
Reproduce analysis	Code, environment file, data dictionary, sample data	Plain-text code and documented dependencies	Open code with restricted or synthetic data
Preserve tabular results	CSV, spreadsheet, statistical package file	CSV with units, labels and missing-value rules	Remove personal or confidential fields
Retain geospatial evidence	GeoJSON, shapefile package, GeoTIFF, map PDF	Document coordinate reference system and precision	Aggregate locations when re-identification is possible
Support controlled reuse	Metadata record, application form, governance notes	Repository access controls and named custodian	Restricted access with review criteria
Connect related outputs	DOI links, publication, grant and session record	Persistent identifiers and version history	Keep corrections and superseded versions visible

Make The Record Findable And Reusable

A deposited file has limited value if nobody can find it. Use a clear title that describes the subject, location and period without relying on an internal project code. Add names in a consistent form, institutional affiliations, ORCID identifiers where available and keywords used by the relevant research community. Spell out acronyms in the abstract and metadata.

Citations should treat the archive as a scholarly output. Include the creator, year, title, version, repository and DOI. If a later article uses the archived dataset, cite both the dataset and the paper when appropriate. This gives credit to data creators and allows repository metrics to reflect reuse.

Versioning is essential when a file is corrected or extended. Do not overwrite a deposited dataset without an audit trail. Explain what changed, why it changed and whether published results are affected. A small correction to a label may need a new version; a major reprocessing may require a separate deposit linked to the original.

Revisit the record periodically. Check the DOI, repository permissions, external links, software references and contact details. A brief annual review is practical for active projects, while a review before staff departure or grant completion can prevent avoidable loss. Australian universities frequently reorganise digital services, so relying on a familiar internal URL is risky unless a persistent public identifier is also available.

Archiving should support reuse without encouraging careless reuse. State the intended scope, known limitations and required acknowledgements. If the work concerns regional communities, Indigenous knowledge, environmental sites or vulnerable populations, explain the conditions under which reuse is respectful and lawful.

Practical Recommendations For Presenters

The best time to prepare an archive is before the conference, but a post-event review can still recover most of the research record. Set aside a defined session to inspect the files, confirm permissions and write the explanatory material while the methods and decisions remain fresh. Treat the archive as part of research communication rather than an administrative afterthought.

For Australian presenters, include local context where it affects interpretation. A dataset collected across Sydney, Melbourne, Perth or remote communities may reflect different transport patterns, service access, climate conditions or digital connectivity. Dates should be unambiguous, measurements should use Australian conventions, and references to legislation, ethics approval and custodianship should match the jurisdiction and institution involved.

• Inventory slides, data, code, documentation, permissions and related publications.
• Remove personal information, hidden comments, passwords and unnecessary working files.
• Create a README, data dictionary, changelog and clear citation instructions.
• Select stable formats, a reputable repository, a licence and an appropriate access level.
• Record Indigenous governance, participant consent, copyright restrictions and data custodianship.
• Test the deposited record from a separate device and verify its persistent identifier.
• Review the archive after corrections, staff changes, publication and major repository updates.