The approach we are taking at the moment is to use either or both of the following:
-
Standard web server statistics, from the Apache web server with a package like AWStats
-
Remote stats services like Google Analytics or W3Counter, which work via JavaScript code embedded in pages on the repository site, or any other website for that matter.
Both these approaches require human interpretation, or some kind of processing or log files and reports as they do not take account of the structure and function of the repository – they will not distinguish between the various kinds of pages and views in the site, but they are all that resources will permit us to use in the first release of the software.
We looked for a way to provide useful repository-specific statistics without a lot of development. Oliver Lucido read up on the PIRUS2 project which looked at repository statistics that could be collected and reported across repositories, but he found that it is concerned with documents, chiefly journal articles, and not suitable for tracking use of research data collections, which is what we’re dealing with here.
Looking to the future we note the following:
-
The Fascinator. which provides the building blocks for ReDBox, has a event logging system, so a future project will be able to add reporting on
-
We would like to work with ANDS on a way to start collecting usage statistics, in research data registries/repositories so a page view of the metadata about a data set in a university repository and one on Research Data Australia for the same thing could be aggregated in one spot; like PIRUS project for research data.
Copyright Peter Sefton, 2010. Licensed under Creative Commons Attribution-Share Alike 2.5 Australia. <http://creativecommons.org/licenses/by-sa/2.5/au/>
This post was written in OpenOffice.org, using templates and tools provided by the Integrated Content Environment project and published to WordPress using The Fascinator.
There’s a rumour that all Collections within the ARDC will be issued with a DOI (by ANDS) and that statistics will be collected somewhere in the DOI resolution infrastructure and made available to partners by ANDS.
See “What is ANDS doing?” in http://www.ands.org.au/guides/doi.html
Hi Conal,
Behind every rumour there’s usually some truth.
As the ANDS guide says, ANDS is establishing a service that would allow Australian research organisations to allocate DOIs to datasets they “publish” (ie make public). ANDS won’t be allocating DOIs, ANDS will be running the infrastructure that would allow research organisations and data publishers to do so if they wish.
The rationale behind the service is to enable acknowledgement of data usage through a more formal citation of the dataset. Again ANDS does not intend to create a new citation indexation service. Using a doi and promoting formal citation would allow cited datasets to play in the existing citation indexing world. It has nothing to do with the DOI resolution infrastructure, which at best might give an idea of hits – but the data publisher would know that number anyway. What the DOI service is targeting is a way to count (or approximate) acknowledged re-use.
We’re very excited about the development, and we plan to have the service in production by mid-year and in pilot very soon. On which more shortly… we will be announcing an early adopters pilot and citation proof-of-concept projects. The scope and policies for the service will also be released over the next month or so.
cheers,
Adrian
Hi,
I can clarify ANDS activities regarding DOIs. ANDS is in the process of creating an optional service for collection owners that will allow them to mint and assign a DOI to a data-set. I’d like to stress that using this service is optional.
We are doing this to try and support a culture of data citation in Australia that parallels the existing culture of publication citation. ANDS is a member of the international DataCite consortium (http://datacite.org) which has been established to support data citation as a peer DOI infrastructure along with Crossref et. al.
If you have any questions regarding this upcoming service or would like to have input into the service development please contact Stuart Hungerford at ANDS (stuart.hungerford@ands.org.au).
Thanks,
Stu
Hi Peter,
interesting post; reminds me of the discontinued BEST project from APSR days.
see http://www.apsr.edu.au/best/index.htm
It’d be interesting to consider opening up Research Data Australia logs to such an aggregation service.
ANDS would have to (you’d think) eat its own dog food and distribute usage data in machine actionable ways…
Adrian