The concern to having a proper RDMP was triggered by difficulties faced by researchers to find data from another researcher or previous research and to extract data from reports. The other problem is to find guidelines, especially in Indonesia, on how to appropriately manage your research data, to store them, and to keep them available in the long run. Clearly scientists have issues on how to re-use dataset from prior research, how to cite them into their own work (re-use), and how do they know the limitation of such action.

Due to the large effort to get data in terms of funding, time, and energy, the life time of data should be more than one or two years, as we find to be the general nature in Indonesia research ecosystem (Irawan 2018, Neylon 2017a, Neylon 2017b, Borghi et al. 2018) (Fig. 1). Another important point to address is the barrier of data sharing that involves the fear of getting scooped, the lack of knowledge to Intellectual Property Rights (IPR) and data ownership. Therefore, by developing this document, we could solve the barriers and at the same time we could come up with another way to increase the value of research data, instead of only looking at mainstream metrics.

Figure 1.  

Current situation of data lifecycle (Irawan 2018).

Researchers can use this article as an assessment form to describe the setting of their research and data management requirements from potential funder. Researchers can develop more detail RDMP to cater specific project's environment. They should justify the setting of their research and requirement of the funder regarding data sharing and data preservation.

The proposed RDMP is divided into seven components:

1. Data collection
2. Documentation and metadata
3. Storage and backup
4. Preservation
5. Sharing and re-use
6. Responsibilities and resources
7. Ethics and legal compliance

Given the different nature of research, funders, and DMP standards, we refer to the following sources in developing this RDMP:

• Data sharing culture (Neylon 2017a, Neylon 2017b)
• Open data principles and reproducible research (Irawan et al. 2017)
• RDMP check lists or rubric (Digital Curation Center 2016, Teperek et al. 2017, University of California Curation Center 2018)
• RDMP case study from various fields of sciences (Neylon 2017c, Traynor 2017, Wael 2017, Woolfrey 2017)

This RDMP covers the following type of data or document which are considered as data source:

• Raw data that may come in the following forms:
  • any field or laboratory measurements collected during in a research
  • any voice recording and its transcript of an interview or any other forms of data collection phase
  • any vector and raster based images
  • any video recording and its text caption of an interview or any other forms of data collection phase
  • survey form responses from participants
  • field notes or laboratory records
• Grant Proposals: funders may request researcher to submit their research plan as pre-registration document in several platform such as OSF or Curate Science.
• Project-level RDMP: some funders, such as RCUK, mandates the submission of final RDMP before the project begins.
• Shared texts, voice, or video recordings of communication between team member.
• Reports: this documents may appear as preliminary report, mid-term report, final report, or short communications.
• Preprints: preprint has been admitted as part of research output by several funders (Bourne et al. 2017).
• Maps

Although most of researchers use Microsoft-based applications and Most open repositories accept and provide native viewer for many formats, but the following are our choice of formats. You may refer to University of Sydney RDMP file formats or Cornell University’s preservation file formats for more information.

Spreadsheets

It should be written in text format, eg: csv (comma separated value), or txt (using tab separated value). Data creators should format the spreadsheet in a "database" format:

• start the data immediately in cell (1,1),
• avoid merging rows or columns, and
• clearly use the correct and consistent cell format, eg: number, string, date, time, category.

Documents

We recommend text-based (ASCII) file, eg: txt, Markdown, or any other text format that can be created and read using plain text reader like Notepad

Audio/video recordings

• Audio recordings: wav or mp3
• Video recordings: mp4 or mpg

Images and maps

• general image: jpg, png, bmp, tiff
• raster: geoTiff
• vector: SHP

Emails (project communications)

Although most researchers are now using proprietary email clients like Ms Outlook or Apple Mail, but they need to store selected emails in to plain text as well.

Files are uploaded to online repository and organized into folders by phase or by working package. If the file organization get too complicated to accommodate into a set of folder structure, then it should be separated and linked together. We recommend the following set of folders to organize the files.

root folder:

• data:
  • raw
  • processed
• analysis
  • code (or script)
  • tables
  • figure or image)
• output
  • report
  • presentation
  • article (or manuscript)

Some field of research may have other specific folder arrangement, but generally they should have the components in the figure. If some team members choose to maintain a Google Drive, DropBox, Onedrive or other cloud services, then they should make an accessible link to the drives or folders and register the links to the data repository. To accommodate limited storage, Principal Investigator (PI), Co-PI, and team members may also maintain an open repository, such as: OSF, Figshare, Zenodo, GitHub, GitLab, and other similar services, given that such services offer version control and access option features. All services should be linked together to a central repository. The team may also maintain a dedicated project website to store the data and related research documents, to keep track of the activities, and to store the project's repository or storage structure.

All data will be preserved in open formats to ensure that its readability in the future. A metadata should be attached into each data file, or in some instance, a data folder. A Readme file should be included in the root folder containing folder structure, general overview and some context of the data.

All deliverables (data, reports, presentations, preprints) should be recorded, listed, and stored in the project repository. A Readme file may be useful to describe the context, time frame, location, structure, and status of the files. A data staff (DS) may be assigned to check the status of the documentation.

We recommend the following minimum metadata schema for general data:

• Title of the dataset (see example)
• Abstract (to give context)
• Creator
• Contributor
• Publisher
• Funder
• Date of publication
• Resource type
• Location
• License/rights
• Data structure
• Data size
• File format

For geospatial dataset, we refer to the ISO 19115-1:2003 geospatial metadata standard, which is also used by the Badan Informasi Geospatial of Indonesia (Indonesia Board of Geospatial Information). The following tables contain minimum metadata schema for general dataset and general geodataset (link to worksheet, open the related sheet).

We anticipate less than five gigabyte of data and documents to be generated by the project. As far as possible data will be deposited in long term archives. A minimum of 10 years of preservation should be in consideration, but there are open repositories that provide longer preservation, eg: up to 50 years or more. Data should be deposited at the start of the project and ended by the time final report submitted to the project funder. An embargo period (maximum of two years) may be assigned if needed. Following the end of the embargo period, an assigned data staff must make the data publicly available until minimum of 10 years.

Data and documents are stored on a three storage levels:

• working offline storage and at least one offline backup using portable hard drive
• online dynamic data repository using: university's available institutional repository and/or open repository services like the OSF (maintained by Center for Open Science), Figshare (maintained by Digital Science), or Zenodo (maintained by CERN).
• online static data repository: institutional repository can be used to store the final dataset and other documents.

We suggest the following backup strategies:

• backup from offline working storage to portable media must be preformed immediately, daily backup is highly recommended.
• backup to cloud storage or repository at least once a week.
• team members are suggested to use backup application such as Apple Time Machine or Free File Sync.

The research team, relevant members of the research team, and project participants will be granted access to the data repository and to other online services. The access will be set through a unique userid and password system before embargo period ends. The minimum access for the above-mentioned parties will be "read-write" access. While "administrator" role should be given to the PI and at least two other team member one Co-PI and data staff. After exceeding the embargo period, the data repository will be made public.

Selection of material

All final materials as follows will be kept available in the Institutional Repository and OSF dynamic repository:

• data:
  • raw data
  • final processed data
• reports:
  • preliminary report
  • mid term report and
  • final report

All intermediate and ongoing files, including data and other documents will be made available in the OSF dynamic repository.

Preservation

Long term preservation of publicly available data will be through appropriate repositories including institutional repository. More than one archive may be selected using the LOCKSS principle or FAIR principle for data sharing as the main criteria. In this case, we recommend OSF dynamic repository and static institutional repository.

For all data generated from a research, we may ask the data creator to convert it from any proprietary file formats to open formats for long term preservation. Other option would be to have a data staff (DS) may be assigned to work on file conversion. The data creator or DS should ensure the anonymization/de-identification upon sensitive data.

In general sense, we recommend sharing raw, processed, analyzed, and final dataset. However, given the nature of the project, PIs may appeal for another form of data sharing. They could fill in data assessment form in order to come up with appropriate data sharing mechanism. PIs may have to:

• choose which type of data that they think could be safely shared without breaching data release agreement with other parties.
• separate primary from another institution from the primary new data acquired by team members.

We recommend to use moderate licenses, eg: CC-BY license, MIT license, and Academic Free License, as the default license for data and also for all resulting documents. However for PIs may propose another license, such as CC0 waiver, CC-BY-SA. For sensitive data, PIs may suggest a restrictive license.

All data and data repository should be able to be found by at least one indexing services, eg: Google Scholar. Common repositories are now accessible via BASE and ONESearch (a feature from Indonesia National Library and Archive). To be formally cited, we also recommend the usage of persistent link, eg: DOI from CrossRef or Datacite.

PI and an assigned DS are responsible for research data management. This includes file conversion, classifying and managing the various research outputs identified in this RDMP, throughout the research cycle and during the lifetime of the data.

In the case of a change of PI or DS, responsibility will be transfered to the one of the Co-PIs or to a DS assigned by the PI or institution.

Aside to the data collection phase, the major costs of data management for the project are for management and storage components. The management components should be funded by research project, while the storage is the responsible of university, or a PI may select a free-open repositories.

A university-level or several faculty-level Data Steward (DS) should be assigned to ensure the management of sensitive data and general data management in general. The access to the such data may be restricted to PI, one of the Co-PIs, and the DS. The DS will have a checklist form to help them assess the situation.

Users must register to access the data or contact University DS and filling out a sensitive data usage form. The form then will be evaluated by university-level or faculty/school-level DS, given that the DS should also consult with the data creator or original researcher.

IP rights for the project are held by the university, or it could be a joint IPR management for joint research activity. It should be clearly mentioned in the data agreement.