Chatting with @clarileia from the OU Library today about what a) interesting, and b) useful things we might be able to learn from web analytics around the OU’s open repository – http://oro.open.ac.uk – I wondered whether it would be possible to generate reports based around traffic coming from other HEIs.
I had a vague memory of setting up filters on Google Analytics years ago to segment out library activity based on visitor IP address, using IP ranges from from the different OU regional offices to generate reports based on library website usage by region, though I’m not sure I ever blogged it (I was asked not to publish the list of OU IP address ranges…). Trying to refresh my memory, it seems you can set up custom filters in a Google Analytics site profile to limit data collection to visits from within a particular IP address range (eg exclude internal traffic and IP address range tool):
IP limits don’t appear to be available for defining custom segments; instead, GA looks up the owner of the IP address and reports that as a Service Provider attribute, which can be used to define a custom segment:
(When accessing GA though the API, I think the corresponding field is ga:networkLocation(?), though I haven’t tested it…)
Here’s an example of what a segment filter on Service Provider terms containing university turns up:
So now I’m wondering: is there a full list of “Service Provider” names for UK HEIs, as picked up by Google Analytics, anywhere, that could be used as the basis of a shareable/templated Advanced Segment?
See also: OUseful.info posts on library analytics.
OUseful.Info, the blog... 
Not sure this fits the bill. However, it should provide a RESTful list of names:
http://www.jiscmu.ac.uk/news/view/189
http://www.jiscmu.ac.uk/api/
http://www.jiscmu.ac.uk/api/services
Hope this might help. I came across the JISC Monitoring Unit somewhere, so I stuck a page on the XCRI KB: http://www.xcri.co.uk/h2-mm-resources-a-z/124-jisc-monitoring-unit-mapping-of-institutional-identifiers.html.
Alan
@alan I know there are lists of identifiers as eg maintained by JISCMU – I wasn’t sure that IDs Google was using or how it was identifying them? Or whether someone had compiled a list… A principled experiment would be to try to identify IP ranges, set up filters-in around those and then look to see what identifiers turn up for each domain name range. (Is there a list of HEI IP ranges?!)
Not sure how helpful it will be but we’re currently in the process of adding IPv4 addresses to our list of organisation identifiers (look at http://data.jiscmu.ac.uk/rest/organisations/id/709/identifiers for an example) using data from Janet’s network team.
Any information on GA’s list of “service providers” would be really useful
@herrdoktorc thanks – I think there may be a ToS busting route to IP address sampling in Google Analytics, but I’d have to do some digging… Agreed it would be really handy to know where Google get their IP-service provider name mapping from…
@alan I know there are lists of identifiers as eg maintained by JISCMU – I wasn’t sure that IDs Google was using or how it was identifying them? Or whether someone had compiled a list… A principled experiment would be to try to identify IP ranges, set up filters-in around those and then look to see what identifiers turn up for each domain name range. (Is there a list of HEI IP ranges?!)
I had a go at creating a filter on IP too, but couldn’t get it to work and couldn’t work out what the problem was so created one excluding ISP domain instead – tomorrow I will see if it has worked!!
@clari I think I’ve managed the IP range filters before? Were you trying to just exclude OU IP ranges? Bear in mind that the regions may have their own values that return an open university name/registrar/owner.
I copied what you (or someone) had done in another filter in GA but it didn’t like the RegEx – was fine to save the one that was already there but I got an error when creating a new one, even when using the predefined filter and filling in the boxes rather than doing it as a custom one!