The Data Management Plan (DMP) is a formal document that describes how the data and metadata has been obtained, processed, organized, stored, secured, preserved and shared, both during research and after the project has ended1. It is now a prerequisite for all ERC projects benefiting since the Horizon2020 funding campaign (article 29.3), for which there is an obligation to ensure free access to metadata and publications resulting from the funded programs. The DMP must produce (re)findable, accessible and as far as possible open, interoperable, reusable data, well known as the FAIR principles.
Indeed, a DMP is an essential tool both for funders and research organizations (optimization and return on investment based on the re-usability of data and metadata), but also for researchers, because they prove to be effective for the management and backup of the data and metadata, and thus reduce the cost and risk2.
Based on the model proposed by Horizon 2020, the ERC HornEast DMP presented here3 attempts to offer its feedback, as has been recently done with other ERC projects4, as well as highlighting the questions, issues and some good practice proposals for future projects in progress of preparing, writing or rethinking their own DMP.
DMP mid-term review
According to the European Commission’s Guidelines on Data Management in Horizon 2020, the tasks finally carried out to update the first version of the DMP established in July 2018 (submitted in the first 6 months of the project) were:
– implementation of management tools;
– revision of the DMP;
– deposit of data to be shared in a data warehouse;
– deposit of data to be kept on an archiving platform.
First, we arranged for a group meeting between the researchers of the project and the data engineer in charge of the DMP. This is a short appointment (around 30 mins) that lays the main groundwork and answers questions from researchers about the application of DMP to their work. Then, a private meeting (about 2 hours, renewable twice if necessary) with each researcher aims to explain in more depth the role of the DMP, how the data and metadata will be saved and shared. Each researcher presents his data and informs the engineer responsible for the PGD which can be made accessible, under embargo, or must remain private.
Following each individual meeting, the data of each researcher is copied in personal files to an external hard drive. These data are then processed (renaming, cleaning, elimination of duplicates between researchers’ files, mainly photographs taken during surveys campaigns and field excavations), converted to open formats and saved both on an external hard drive and in a platform dedicated to the processing of scientific data: Sharedocs (file manager implemented by the TGIR Huma-Num).
This allows to keep a copy of the original format, and reduce the potential loss of formatting during file conversion. In this way, researchers can both keep their data in the way they filed and named it (and maintain the integrity of a record) and use their personal folders on Sharedocs to continue filing and processing their data. Only in a second step are the files refiled and renamed by the data engineer in charge of the DMP according to an agreed classification (to consult the classification of the data see page 9 of the DMP available down below).
This ensures both transparency and monitoring of data processing for researchers and the data engineer in charge of the DMP.
Issues and ethicals questions
One of the most common questions researchers ask during their interview is which data (and which versions) should I deposit? Indeed, research data is defined as factual records (numbers, text, images and sounds), which are used as primary sources for scientific research and are generally recognized by the scientific community as necessary to validate the results of research5.
The ERC Horn East Project collects and generates its data through:
– Field observations, archaeological surveys and excavations, survey results, interview recordings, photographs;
– Research in various sources (chronicles, biographical data, funerary inscriptions, collections of artifacts, etc.);
– Laboratory equipment (as C14 dating) or geographic software.
It is therefore data that is both raw data, derived data and data resulting from analyses.
However, the work process of some researchers leads to using several formats before arriving at the final “product”. For example, an epigraphist working on inscriptions which are mainly recorded through photographs taken during field missions (.jpeg), produces facsimiles with graphic software (.ai), which is then copied into a text file (.docx), to end up in a publication (.pdf). In the same way, an archaeologist working on his field surveys report digitizes a sketch (paper to .jpeg) which is also taken up in a graphic software (.ai) to be reworked and subsequently inserted (.jpeg, .tiff) into a finalreport or publication (docx. to .pdf). This “intermediate” data sometimes represents several gigabytes, and is oftenaccompanied by notes and informations that do not appear in the final publications but represents the research process, the way in which the researcher thinks and works, and can therefore be considered as a precious archive per se.
A second interesting issue is the question of file renaming. The project has created many geographic information system (GIS) files, however, the issue of file naming not having been established upstream, the files are named with spaces and special characters, two specificities that hinder both the interoperability of the document and its long-term archiving6. However, renaming them would cause the path of the file to be lost in the software, and thus complicate its reuse.
From an ethical point of view, the question concerning the choice of the deposit of the data has also raised the attention of several researchers who are already working on certain research subject before intervening in the project. Indeed, the specificities of projects as the ERC (European Research Council) or ANR (National Research Agency) are to be designed by various, multidisciplinary researchers. However, some of these researchers can also take part simultaneously in other projects, or have been funded only during part of the project (months, years…), which often explains the important turn over characterizing this type of projects7.
Consequently, some researchers may be reluctant to share their data which they still consider as unfinished work. It is then important to make the distinction between the data produced by and within the framework of the project. For example, it can be a conference presentation given by the researchers during a scientific event, or a publication for which the researcher was funded (mainly completed work), while the raw data (on which the conference presentation was partially based) collected over several years by the researcher and which is still in the process of analysis, is not to be considered as data produced uniquely by the project, and there for be deposit and shared.
In any case, it is important to remind researchers, whether their data will be deposited, accessible in open access or under embargo, that their copyright will always be maintained (intellectual property rights). This is why it is also important to raise awareness about the different licenses option available when data are sent online.
Preliminary assessment and future prospects
To date, nearly 130 GB of data and approximately 2 linear meters of standard format archives have been produced by the project (to consult the description and the classification of the data see page 9 of the DMP available down below).
Out of a total of 25 publications published so far (journal articles, book chapters, reports, blog posts) deposited on HAL, 23 are accessible in pre-print and original version, which corresponds to 92% of the open access publication (see table below). Out of a total of 283 photographs (of 6073) and maps deposited on MédiHAL, 264 are accessible without embargo, which corresponds to 93% of the images and maps open in open access8.
Table: ERC HornEast publication deposit in HAL
A total of 22 datasets have been deposited in Nakala (representing approximately 40 GB of data), classify in 7 collections:
Although any ERC project is required to share it’s publications and metadata, in the case of the ERC project HornEast, two aspects made this requirement critical.
First of all, the political evolution in Ethiopia during the frame work of the project and the war that erupted in the region under investigation increase the value of the collected data. Many of the sites documented during the project are now inaccessible and/or potentially damaged, therefore sharing the data (even unpublished) will benefit both to local and international community.
Secondly, as archaeology is by essence a destructive process, making available the raw data produced from excavations, ensures the knowledge on heritage is not lost and can be spread among the scholarly community.
Good practice guide for DMP
Below are presented some recommendations raised by the project team during a meeting dedicated to the project archives on April 25, 2022:
Collection of the data in a shared space (hard drive or online storage)
– if this is not done systematically throughout the project, try to schedule a day every three months to gather the data. Grouping the data upstream of the classification facilitates both file renaming and processing/cleaning (ensuring that there are no duplicates). Also, presenting the data collected so far to the other members of the project offers the possibility of communication between the researchers and can sometimes lead to new reflections or common lines of work.
– if possible, appoint one of the researcher “responsible for data processing” who will accompany the project throughout. If a specialist in scientific data processing cannot be attached to the project, try to establish a link with the staff of university libraries, or the sector responsible for the digital humanities of your establishment to help with the drafting of the DMP from the start of the project.
Indeed, the editor of the DMP must wear several hats, as it is necessary to be familiar with several aspects of data processing: the quality of metadata (the importance of communication with researchers to provide as much information as possible must be noted), knowledge of storage platforms, warehouse and standards (a specialist in scientific information), long-term conservation (archivist) and masters the obligations to be respected towards intellectual property, the confidentiality of (meta)data (jurist).
– try to establish a file naming standard as early as possible in the project.
– when dealing with photographs (or other data) taken during field missions, before the researchers rename the files (for example “fig. 3” to associate it with a publication), ensure that the photographs of the researchers have been grouped and copied into a folder or external hard disk to simplify the verification of duplicates between researchers.
– Try to rename the files with the most precise typology. For example, if it is a publication, rather than indicating “publication” in the title, put what type of publication “article”, “book chapter” etc. accompanied by the name of the author and if possible the first words of the title. This makes it possible to offer a more precise description of the data and thus, if it is deposited in a data warehouse (such as Nakala) it is possible for the user to identify the data more easily.
Why? Nakala, while allowing the deposit of datasets, does not offer the possibility of creating a tree structure in the collections (which can impact the granularity of the metadata). In fact, to maintain consistency, the project has decided to deposit all the photographs of a mission in the same dataset, the documentation of a year’s excavations, etc. The distinction of places or type of document in the title thus allows users to quickly identify the data they are interested in.
Enrich the metadata
– if a person outside the project is responsible for the description of the metadata, it is recommended to rely on the reports and publications produced by the members of the project to supply all the information necessary and thus further enrich the metadata (this also makes it possible to verify and not to interrupt too often the researchers in their work).
Ethical and legals aspect
– Remember that if a researcher uses an image that is not of public domain or for which he has not received authorization (for a slide show at a conference for example), it cannot be shared in OA, moreover, without mentioning the source of the image. The same applies if a researcher uses an image taken during field missions, but where children appear.
Access and sharing of the data
– Use a controlled vocabulary for publishing metadata. The scientific communities can then use tools such as opentheso or PACTOLS for archeology to meet standardization needs.
Consult the ERC HornEast mid-term DMP
Consult the data classification scheme
Consult the paper archives classification scheme
The final version of the DMP will be updated at the end of the project.
Written by Maryasha Barbé.
- European Commission. Guidelines on Open Access to Scientific Publications and Research Data in Horizon 2020 (Dec. 2013), p. 10 [↩]
- A recent study show that more than 53% find the DMP to be useful beyond it being a requirement from the side of the European Commission, see Daniel Spichtinger, « Data Management Plans in Horizon 2020: what beneficiaries think and what we can learn from their experience », Open Res Europe NaN, 1:42. [↩]
- The DMP is a living document throughout the life cycle of the project. Specific updates and deliverables must be defined. The DMP presented in this blog corresponds to the intermediate PMD (mid-term). Therefore it will be updated (mostly in terms of data volume) at the end of the project. [↩]
- The ERC Graph-East Project offers in a blog post written by Manon Durier their own feedback on the writing of a DMP [↩]
- Translated from French to English by the writer, Principes et lignes directrices de l’OCDE pour l’accès aux données de la recherche financée sur fonds publics, p. 18 [Online : https://www.oecd.org/fr/science/inno/38500823.pdf]. [↩]
- For some indication on how to name your data, see : https://doranum.fr/stockage-archivage/comment-nommer-fichiers/ [↩]
- A recent publication of a Vademecum for the reusability of data also raises this issue, see: Julie Aucagne, Marguerite Bordry, Camille Desiles, Francine Filoche, Anne Garcia-Fernandez, et al.. Vademecum pour la réutilisabilité des données : Groupe de travail Réutilisatibilité, Consortium Cahier. p.7 [Rapport de recherche] Consortium CAHIER – Huma-Num. 2022. hal-03630095. [↩]
- 283 images of the project’s excavations and surveys in the Tigray region have already been uploaded to MediHAL. [↩]