eDate

08 Jun 2020

Connection Link

https://iu.zoom.us/j/661265362

Attendees

• Jon Cameron

• Randy Kuehn (Louisville)

• Robin Pike (UMD)
• Jon Dunn

Agenda

• Analytics in Avalon

  • How do you use analytics on your sites? Has it been useful or frustrating? Is there anything the Avalon team could do to make analytics easier or more helpful? Attendees are invited to share and discuss how they use analytics in Avalon or other institutional sites or services.

Meeting Notes

IU uses Google Analytics, we'd like to use Matomo.

Randy doesn't have it set up, but the plan is to use Google Analytics. They hired someone to do stats for the libraries, and that person is putting together a data warehouse of sorts. They're looking for anything that can integrate with data warehouse type stuff.

Robin: Not sure what Maryland is using, using GA in their current repository. Will probably use Google Analytics. With GA right now, it's great at providing item stats but not at the collection level.

JWD: David Schober suggests an approach of making a change to the code that configures the web page code, and then one would be able to filter within the strings within the title. How can we send additional information to Matomo that will help send additional metadata?

Infrastructure: at IU, we don't have a data warehouse for the library. IU as a whole has moved past the data warehouse as a concept to focus more on the reporting dashboards. ds.iu.edu has all of the reports and dashboards, and Jon's dream is that we have this data in Matomo, and write some code to pull the data from Matomo using Denodo, to Tableau. We're working on that for a lot of physical library stats, in terms of pulling counter data. The next step is to get our repository usage.

Denodo is not open source. It's a  commercial tool which IU has site-licensed.

Randy would be interested in testing out Matomo. He hasn't done anything with it so far. Most of everything he deals with now is based on Rails.

Would it be worthwhile to have a dashboard within Avalon? Probably not. Randy and Robin are of the same mindset, that it wouldn't be super high priority. Especially when looking at usage and downloads. Anyone with Analytics access would already be able to get what they need anyway.

Robin has done a number of crowd-sourced fundraising projects, and choosing to digitize similar things based on what's getting used. Hard to tell right now, so using reference statistics with Aeon instead, instead of the digital collections repository. Pursue grants or crowd-sourced funding, based on what's high use in the repository. Collection-level information, or even collections that are grouped (by unit). Interested in raw metrics, like pageviews. Can't tell if they're watching the content. Can't tell if they're downloading the file, etc.

Length of time on page would be useful.

JWD: a lot of teaching and learning videos hit the stats of: how many are playing through, how much did they watch and what parts.

Robin: Honestly, I don't care if they're watching the entire time, or if they're skipping ahead. A lot of what's in there is being used in Canvas, it's all bundled together. Would be useful to know if they click on the file to watch it. Or they're looking for something, and accidentally go AV when they should've gone image.

Jon mentions two things in the Hyrax community: Oregon State, adding better analytics to Hyrax. Another is a grant from Arcadia, working with Ubiquity press and to add Analytics to Hyku.