Errors make half of public datasets unusable

A recent study of ecology publications showed that accompanying datasets are often incomplete or of such low quality that the data can be difficult or impossible to re-use.

The study surveyed the datasets accompanying 100 publications and found that 56% were incomplete, and 64% were archived in a way that partially or entirely prevented reuse. The authors listed a number of key recommendations that could help make the data more useful. These include:

• Planning for the management of data at the start of the project
• Using established repositories
• Provide information about the data (metadata)
• Use descriptive file names
• Share raw data if possible
• Use standard formats
• Facilitate data aggregation – through standards and appropriate public databases
• Perform quality control
• Choose a publishing license
• Decide on an embargo

And in a nice touch, the authors of the publication have released the data on the repository figshare: http://dx.doi.org/10.6084/m9.figshare.1393269

If you require assistance with any aspect of data management, please contact the library digital scholarship team at digischol-library@unimelb.edu.au.