Exploiting institutional activity data
JISC is currently funding a range of projects that investigate how data stored in institutional systems can be mined to gain insights into the way that university services are operating and use those insights to improve the services. These projects are spread across two programmes, the activity data programme and the business intelligence programme. There are a few other projects working in similar areas spread across other programmes.
Last week we took the opportunity to bring most of these projects together to discuss their various approaches and to think about what else JISC can do to help universities make the most of their activity data.
We started the event with lightning talks from each project attending. I suspect faithfully listing all of those projects here would probably make for a gruelling reading experience. So instead I’ll group them into the broad motivations the projects are pursuing. Some projects fall under more than one category. I have linked to the presentations the projects gave, if I do not have the slides I have linked to the project website.
Some are mining data to gain insights into behaviour of people or systems in the institution to allow better resource allocation and intervention at crucial periods
LIDP, Supporting institutional decision making, Bringing corporate data to life, Lumis, IN-GRiD, Retain, Student Engagement Traffic Lighting
The projects cover a vast range of areas from libraries to student management to environmental monitoring. Despite this breadth there are some common issues. These are the issues that jumped out at me on the day:
- Not all institutions have people with the technical skills and the statistical skills required to manipulate and analyse data .
- A lot of these datasets are large and that brings up issues of how to store and manipulate the data and how do you decide what to retain.
- Institutions might need to take an institution wide strategic approach to deciding what data should be collected, how it should be exploited and by who. There also needs to be long term strategic approaches to data exploitation in departments like libraries.
- Working across different silos of activity data is a problem we are only just beginning to face.
- Language is an issue. Throughout this blog post I have used the term activity data, but this is a generic term we need to be clearer about what type of data we are talking about.
To help ensure that others can benefit from the lessons the projects learn there are synthesis projects for the activity data programme and the business intelligence programme. The purpose of the synthesis projects is to gather information on key issues and turn them into advice and guidance than anyone in the HE sector can use to inform the way they do things at their institution. Infonet are producing a business intelligence infokit for their synthesis project. The university of Manchester and Sero Consulting are producing the activity data synthesis. You can read the activity data synthesis blog which talks about progress so far and can also see a mindmap that describes the areas their final website will cover.
We ended the day with a discussion of the possible ways that JISC could look to address some of these issues. We produced a long list of very good ideas. We also had a go at prioritising which were the most pressing or valuable. Our top 6 ideas were:
- Developing guidance for institutions on taking a strategic approach to exploiting activity data
- Addressing the need for new skills for exploiting activity data both from a technical perspective and from a statistical skills perspective
- Establish clear definitions for terms used – this could include a simple glossary and use of examples to illustrate the terms
- Developing a culture of exploiting data in institutions
- Exploring what’s involved in ensuring data is easy to reuse
- Study behaviour and how it relates to usage patters
JISC won’t be able to address all of these issues straight away, so my colleague Myles Danson and I will have to decide which we focus on. Comments and advice would be very welcome indeed!
After the meeting Mathieu from the UCIAD project wrote a very interesting blog post about the need to take a user centered approach to activity data so that’s something we’ll need to consider too.