Seven years on from updates to our data sharing policies, we reflect on data-deposition-related changes to our published papers and continue to strongly encourage authors to make source data openly available.
Since late 2018, all published primary research articles at Nature Materials have included a data availability statement (DAS)1. In addition, as part of a broader initiative at Springer Nature to encourage the public sharing of data, we instigated a request to authors to either deposit relevant data in public repositories or to include data with the HTML version of the article. As we highlighted in an Editorial in 2019, authors of accepted articles are asked by the editor to provide ‘source data’ for the figures within the manuscript and the Extended Data figures2.
Credit: Andriy Onufriyenko/ Getty
More broadly, we ask authors to provide the ‘minimum dataset’ that is necessary to interpret, verify and extend the research in the article. If this minimum dataset includes data not directly presenting evidence for a figure, these data files can be provided alongside other supplementary information; however, we have a strong preference for large datasets to be made available in repositories rather than as supplementary information.
If data can only be shared ‘on request’ from the authors, we request that brief explanations are given in the DAS, and in correspondence with the editor, for each data type this is applicable to. Possible reasons include privacy requirements or third-party control. In addition, if data are deposited in a repository, the associated unique digital identifier must be cited in the paper.
Now that these data sharing policies have been in place for several years, we have assessed how manuscripts published in Nature Materials in 2024 have fared in terms of the degree to which data have been shared and also how authors have opted to share these data.
Of the 172 Articles and Letters published online in Nature Materials in 2024, about 3 out of 4 manuscripts include source data either integrated in the HTML version of the paper or in a discipline-specific or general public repository. The data associated with about one-third of manuscripts were uploaded to a repository, with Zenodo as the most popular repository, followed by Figshare. In 2023, the Figshare repository was integrated with the Nature Materials submission system, making it possible for authors to upload data to Figshare at submission, at subsequent revision stages or at accept. Although this repository is only one of many that authors can choose (see a list of repositories recommended by Springer Nature), Figshare integration with our manuscript system enables the data — by means of a private link — to be assessed by peer reviewers, and if the manuscript is published, the data are made openly available.
More general analyses of how data sharing has changed over time have been carried out as a collaboration between Figshare, Digital Science and Springer Nature. These analyses probe how the behaviour of researchers has evolved, with an interest in identifying policies and interventions that have a positive effect on data sharing and that have culminated in annual reports, with the ninth report, the State of Open Data 2024, published recently3. Part of the report includes the analysis of the DAS in published manuscripts, and an observation is that there is a positive trend in open science practices, with “modest increases in repository sharing and reductions in ‘on request’”. However, the report also notes that reductions in “on request” and “in manuscript” sharing are typically not accompanied by an increase in the sharing of data in open repositories. In some cases, it is observed that data are declared as “not applicable” or shared through non-repository sites. Another noteworthy conclusion of the report is that researchers are of the opinion that, in general, they do not receive enough credit for sharing their data.
Going forward, although providing source data and openly sharing data are not mandated in Nature Materials, we will continue to encourage authors to do so by means of requests at the accept stage of the article. It should be noted that some data types are mandated (see list) for deposition in a discipline-specific repository for published articles in all Nature Portfolio journals. For example, crystallographic data for small molecules, protein structures and genomic data are mandated for deposition. In the case of crystallographic data, we ask that the crystal structures reported in the article are identified by their Cambridge Crystallographic Data Centre (CCDC) number in the DAS.
Now, more than ever, with an increasing focus on data-driven approaches and the use of artificial intelligence tools in scientific research, our community requires data to be openly available and in interoperable and readable formats to advance materials design at pace. However, the time and effort needed for researchers to appropriately format data should not be underestimated and we acknowledge these efforts made by our authors. With the increasing value placed on the reusability of data, we hope that those researchers who take the time to deposit data are rewarded with credit for their willingness to openly share.
