RDMF12: notes from breakout group 4 (Tracking Uses)

RDMF12: break-out group on Tracking Usage of Research Data

The topic of the break-out group was "Tracking uses: who uses/reuses data (in academia and industry), and what do they use it for?"

• We discussed why we as institutions providing access to research data, are interested in this topic. We came up with the following reasons:
• Promotion of the research outcomes and the institution supporting it.  If there is a cool outcome of data re-use we want to remind everyone that it was our researchers who generated it in the first place. ☺
• Identification of potential collaborators.  If other researchers are interested enough to re-use data, perhaps they are interested in the same questions and more synergy can be generated by working together?
• Special Collections identification.  If a certain set of data is very popular, then perhaps it has extra value and should be publicised as a special collection.
• Ensuring that the right data is kept and that mandates which use the “last-used” date as a starting point can be accurately tracked.
• Validation of the service.  If the data can be found, downloaded and then re-used the service providing these facilities can get credit for the work providing the infrastructure and the value of the content.

The discussion moved onto ways that this can be measured.  It can be summarised as difficult and requiring culture change.  Suggestions were:

• Data citation. A new and promising development, but more likely to capture academic rather than commercial usage.
• Terms and conditions requesting feedback on use.  Easy to put in place, harder to get end-users to comply
• Does a download equate to actual usage?  A discussion already well-trodden in publication circles!
• Doing surveys of registered users or institutional staff. This is labour intensive, but is similar to the effort need for case studies for the REF.
• Finding incentives to report sharing  such as prizes for data re-use.
• There was some scepticism expressed about the use of altmetrics. “Cheap but fluff” was the consensus view.  

We also considered how end-users know that the data they have found is useful and there was some consensus of using visualisation techniques to show a snippet of the data before enabling the whole dataset to be downloaded. So that a little is easy to get to, but more is slightly harder and might include some sort of registration to enable usage to be tracked.

The conclusion was that it is important to be able to track usage, but not all usage has the same value or impact and this part is very hard to establish and it is easier to measure something if you know from the start that it needs to be measured.  As a starting point we agree to start sharing how our institutions capture information about usage.

(Catherine Jones, 24th Nov)