NC3Rs: National Centre for the Replacement Refinement & Reduction of Animals in Research

DRIVER recommendations – Data availability and presentation

How to share and visualise experimental data transparently.

Content

Item 6: Data availability and presentation

Decisions on whether and how experimental data are made available have a significant impact on the utility of a study. In addition, the way data are visualised and presented strongly influences how results are interpreted. Defining the expected output and how best to visualise data during the design phase of an experiment supports open and transparent data sharing and helps ensure that experimental outcomes are clearly communicated when results are reported.

You can find helpful tips and additional context in the grey boxes in each section below.

Design recommendation

Establish how data will be shared and presented transparently.

Plan how data will be managed and shared

Once an experiment has been designed and it is clear what output data will be generated, decisions on how data will be processed, stored and shared can be made. Proactively curating data as it is collected ensures that it is well-organised and accompanied by appropriate metadata, making it easier to share when the time comes to publish. Implementing a data management plan early in the research process can help establish consistent practices for data handling. It also serves as a valuable reference for collaborators or future users of the dataset, guiding them on how the data was processed, structured and prepared for sharing. Key considerations include:

  • Whether raw data can be shared directly or processing is required (e.g. cleaning, normalisation, aggregation, or anonymisation).
  • What types and formats of data will be generated, and whether these formats are interoperable, standardised and suitable for long-term preservation and reuse.
  • How sensitive or confidential data will be handled, including the need for de-identification, access controls or data use agreements. Ethical approval and consent may be required – this should be in place before any data are collected.
  • Where and how data will be stored during the project, and which repositories or platforms will be used for longer-term archiving and sharing. For more information on data sharing and types of repositories see the interactive content (i. Data sharing).
  • What licensing or access conditions will apply, ensuring that reuse conditions are clear and proportionate to any ethical, legal or commercial constraints.

Metadata (additional information that describes the properties, characteristics and attributes of a dataset) can be compiled and curated as the study progresses. This includes details of the experimental design, methods applied, calibration and setting of equipment, so that it can be easily shared once the experiments are complete. 

The FAIR data principles (see Reporting) should be followed when producing, storing and sharing data to ensure data can be viewed, reanalysed and reused without restriction.

For more information on different types of data and data sharing methods see the interactive content (i. Data sharing).

Many general repositories will allow you to begin uploading and curating your data prior to assigning a DOI and making the data live. Curating data into a repository as they are collected can save time and effort during the publication process and help ensure that data and associated metadata are organised and ready for sharing.

Consider how results will be presented and visualised.

Decisions about how best to visualise data can be made once the design of the study is finalised, and the experimental and biological units have been identified. Because the sample size is the number of experimental units per group, plotting data at that level (i.e. one experimental unit = one data point) helps ensure that the figures accurately represent the underlying data. 

For studies that involve sub-sampling, careful consideration should be given to how the data are presented. Plotting individual measurements as separate data points may give the impression that the analysis was performed on the sub-samples which would constitute pseudoreplication. See Item 1: Experimental unit for more information on experimental units and pseudoreplication.

For studies that include multiple biological units (e.g. samples from different patients or cell lines), consider whether the results should be presented together or separately. Combining data from multiple biological units on the same figure can help demonstrate the consistency and generalisability of findings – for example, showing similar observations across multiple cell lines within a single figure.

When selecting a graph type, consider the number of data points to be plotted and the variability within and between groups. Smaller datasets may require visualisations that display individual values clearly, whereas larger datasets may benefit from approaches that balance transparency with readability. In all cases, the chosen visualisation should explicitly convey the spread and distribution of the underlying data, for example with histograms, scatter plots or box plots.

If imaging data are collected, any image processing, manipulation or enhancement should be documented at the time it is performed so that it can be reported accurately. Original, unedited images should also be retained.

Reporting recommendation

Share and present data transparently.

Make data available for verification and reuse

Guidance and standards are available on how, when and in what format data should be shared. When it comes to sharing data, the FAIR guiding principles should be applied:

  • Findable: Metadata and data should be easy to find for both humans and computers.
  • Accessible: Users should know how they can access the data.
  • Interoperable: Data need to interoperate with applications or workflows for analysis, storage or processing and can be integrated with other data.
  • Reusable: Metadata and data should be well-described so that they can be replicated and/or combined in different settings.

A range of general and specialist repositories are available to allow data to be stored and shared. Specialist repositories (e.g. GenBank) may perform additional integrity checks on data and provide additional metadata and should be considered where relevant. 

Providing a clear data availability statement within the article including links, accession numbers or DOIs, a list of data files and data licence can aid in allowing readers to easily access a manuscript's underlying data. A comprehensive data availability statement should include:

  • What data is provided.
  • Where it can be accessed.
  • Under what conditions the data can be accessed and used.

If for some reason data cannot be shared openly (e.g. proprietary, or data files are too large) then any intermediate data available should be provided. This should include a full explanation of the restrictions on the data, how readers can apply for access and under what conditions access will be granted should be included in the data availability statement.

Raw data files should be stored and shared, with any transformations, adjustments or manipulations clearly described. This is particularly relevant for imaging data, where adjustments to make images clearer can lead to a misleading representation of the underlying results. 

Sharing data through structured repositories is widely regarded as best practice and supports compliance with the FAIR principles. For information on other data sharing methods see the interactive content (i. Data sharing).

Balancing utility and practicality

Data can be used for a range of different purposes, these can generally be grouped under:

  • Verification: Using the data to check the results presented in the manuscript.
  • Reanalysis: Re-running the analysis described in the paper to verify statistical results.
  • Replication: Using the data to guide and support a replication of the experiment.
  • Reuse: Using the data for a different purpose or analysis.

When deciding how to share different datasets, consider both:

  • Utility: How likely is the data to be reused or repurposed by others?
  • Practicality: How easy is it to prepare and share the data in a usable format?

The table below outlines common types of data generated in in vitro experiments and key considerations for balancing their scientific value with the practically of sharing them in line with the FAIR data principles. Data types have been categorised based on their potential application, i.e. how other researchers might use them once they are published.

ApplicationsExamplesUtilitySharing recommendation
Reuse or repurpose the data.Omics data (e.g. transcriptomics, proteomics), high-content screening.High – can be reanalysed, used in meta-analyses, or for new hypotheses.Always share openly in discipline-specific repositories. Include full metadata and both raw and processed data.
Replicate findings.Cell viability readouts, Ct values, raw ELISA readings, flow cytometry files.Medium to High – enables replication of findings.Share openly if feasible. If not, share a representative subset with full metadata.
Replicate analysis.Raw image files, time-course measurements, instrument logs.Medium – useful for verifying or replicating analyses.Share openly if easy. Otherwise, share a subset or processed version with clear documentation.
Verify findings.Representative microscope images, summary tables.Low to Medium – supports visual or qualitative verification.Share if easy. Otherwise, include in supplementary files or make available upon request.

For more information on practical data sharing see the interactive content (i. Data sharing).

When assessing the practicality of sharing certain datasets, the monetary and environmental costs should be considered. Some repositories may charge for their services or apply additional fees for hosting large datasets. Additionally, the energy demands and associated environmental impact of long-term data storage in data centres should be weighed against the potential utility of the data.

Present experimental data transparently

Data visualisations should provide readers with all the information required to understand the study design and data produced. Key features include:

  • Inclusion of all data points or methods that allow the distribution to be seen.
  • Clear labelling of axes and groups.
  • Inclusion of measures of variability.
  • Clear description of the experimental set up in the figure legend.

It may not be practical to present individual data points when large datasets are plotted. In this case, the use of visualisations that display the full degree of statistical variation in grouped data, such as box-and-whisker or violin plots, are useful ways to present data transparently.

For imaging data, it should be clear if images are representative and any manipulations should be transparently reported alongside the figures.

While bar charts are a common way to visualise data, they are often used inappropriately to present continuous biological measurements when bar charts are intended for categorical comparisons using summary statistics. Bars can be visually misleading, they do not show the distribution of the data, they encourage comparisons relative to zero and they give the impression that data occupy the full area of the bar and that none exist beyond its height. 

How bar charts can hide key data characteristics is explored in more detail in the interactive content (ii. Presenting your data).

  1. National Institutes of Health (2020). Final NIH Policy for Data Management and Sharing.
  2. UKRI (2018). Guidance on best practice in the management of research data.
  3. UKRI (2022). Open Research.
  4. National Institutes of Health (2026). Data Sharing Approaches.
  5. Nature Portfolio (2026). Reporting standards and availability of data, materials, code and protocols.
  6. PLOS ONE (2026). Data Availability.
  7. Starr J et al. (2015). Achieving human and machine accessibility of cited data in scholarly publications. PeerJ Computer Science 1: e1. doi: 10.7717/peerj-cs.1
  8. Hewitt JA et al. (2017). Accelerating biomedical discoveries through rigor and transparency. ILAR Journal 58(1): 115-128. doi: 10.1093/ilar/ilx011
  9. Miyakawa T (2020). No raw data, no science: another possible source of the reproducibility crisis. Molecular Brain 13(1): 24. doi: 10.1186/s13041-020-0552-2
  10. Data Citation Synthesis Group (2014). Joint Declaration of Data Citation Principles. FORCE11. doi: 10.25490/a97f-egyk
  11. Stieglitz S et al. (2020). When are researchers willing to share their data? Impacts of values and uncertainty on open data in academia. PLOS ONE 15(7): e0234172. doi: 10.1371/journal.pone.0234172
  12. Colavizza G et al. (2020). The citation advantage of linking publications to research data. PLOS ONE 15(4): e0230416. doi: 10.1371/journal.pone.0230416
  13. Piwowar HA et al. (2007). Sharing detailed research data is associated with increased citation rate. PLOS ONE 2(3): e308. doi: 10.1371/journal.pone.0000308
  14. Tenopir C et al. (2011). Data sharing by scientists: practices and perceptions. PLOS ONE 6(6): e21101. doi: 10.1371/journal.pone.0021101
  15. Wilson SL et al. (2021). Sharing biological data: why, when, and how. FEBS Letters 595(7): 847-863. doi: 10.1002/1873-3468.14067
  16. National Institutes of Health (2026). Selecting a Data Repository.
  17. National Institutes of Health (2026). Repositories for Sharing Scientific Data.
  18. Nature (2026). Data Repository Guidance.
  19. Registry of Research Data Repositories (Re3data) (2026). Registry.
  20. FAIRsharing (2026). Repositories.
  21. Taylor & Francis (2026). Understanding and using Data Repositories.
  22. Hrynaszkiewicz I et al. (2020). Developing a research data policy framework for all journals and publishers. Data Science Journal 19. doi: 10.5334/dsj-2020-005
  23. Wilkinson MD et al. (2016). The FAIR guiding principles for scientific data management and stewardship. Scientific Data 3(1): 160018. doi: 10.1038/sdata.2016.18
  24. Tedersoo L et al. (2021). Data sharing practices and data availability upon request differ across scientific disciplines. Scientific Data 8(192) doi: 10.1038/s41597-021-00981-0
The DRIVER recommendations

A set of six recommendations tailored to the design and reporting of in vitro experiments. Find out more on the landing page.

Logo for the DRIVER recommendations
Diagram of six experimental units represented by test tubes with different shaded sections representing different experimental groups.

Go to Item 1: Experimental unit.

Visual representation of risks of bias using two configuration of wells in a plate format. Both groups have three groups represented by white, light blue and dark blue wells. One configuration has groups aligned in the same column, the second configuration has the groups randomised across the plate.

Go to Item 2: Risk of bias.

Graphic illustrating experimental model factors with a microfluidic device in the centre and icons for temperature, CO₂ level, time, and humidity arranged around it.

Go to Item 3: Experimental model.

Visual representation of experimental procedures using a multiwell plate with a pipette filling wells.

Go to Item 4: Experimental procedures.

Visual representation of experimental groups and exclusions using box plots and outlier markers to indicate excluded data points.

Go to Item 5: Experimental groups and exclusions.

Image representing Data availability and presentation featuring a laptop screen with a data folder and an arrow cursor, symbolizing access and sharing of data.

You are on this page.

For any queries or feedback please contact driver@nc3rs.org.uk