We have the statistics that come with the DSpace distribution setup for our IR at Georgia Tech and are looking for ways to get more information out of it. In particular I am interested in getting information per item, possibly bitstream, in terms of how often it was viewed/downloaded. In addition it would be interesting to see which collections/communities generate the most interest. And of cause filtering access from IR administrators/robots or even from within our library would be very useful.
I believe providing individual view/download stats (how often. from where) could be a BIG selling point for researchers/departments. This is info they do not usually get when they put their docs on their web sites and certainly is something one would like to brag about if the numbers look good.
I tried to install the MinHo package (unsuccessfully, I make a separate post for that) and had a brief look at the package from ANU.
Can anybody give a quick opinion on the difference between or similarities of these packages ?
I understand that both packages keep a record of
transactions/interactions with a dspace instance inside the instance's
database. Our instance receives quite a bit of traffic, this month we
already had 140,000 itemviews and 100,00 bitstreams views. Collection,
community views and Searches are all comparably few ranging between
7,000 to 16,000 this month.
- How much do I need to worry about the database growing and growing and growing ?
- Is there a facility in the stats package that purges old log records ?
- Is there a way to request holding aggregate information instead of detailed interaction info ?
- I understand that statistics are generated on the fly on the basis of
up-to-date info from the database. If that is true: How does that impact response time
when generating stats-pages ?
Georgia Institute of Technology
Library and Information Center