Investing in Business Intelligence

There are many areas of the university already using Tableau to visualise data; many people have been trained to build Tableau visualisations and receive support from CIS as they develop their ideas.  The use of Tableau continues to grow organically.

The need for data visualisation cuts across many different initiatives of Digital Lancaster

Investing in Business Intelligence means investing in systems that let us:

  • Understand how things are now.  Whether that means how our finances are, how many students we have, how our research income looks, we want to know where we stand.  And data visualisation can be much easier to understand than tables of numbers.
  • Understand how data relate to each other.  Perhaps we need to combine data from LUSI and Agresso, so we can understand which PhD students in our department have and haven’t paid their fees.  Or perhaps we want to relate data in Pure to data in Agresso to understand the research income from each department.  But unless we relate data from one system to that in another, we won’t get a complete picture.
  • Understand how data is changing over time.  Sometimes what we really want to know is how where we are now, compared to where we were this time last year.  It can be hard wind back the clock to have create that view.  It can be much easier to take a snapshot of the data periodically, and then compare those snapshots over time.

Most of the visualisations being created at the moment fall into the first category.  And such visualisations are by far the easiest to create.  No snapshots are required, nor do we have to marry up data from separate systems.  There is plenty of value in such systems, and that’s the feedback we’re getting from the people creating and using them.

However, there are also needs which may only be filled if we relate data from separate systems (or silos) of data; and still other needs which can only be filled when we can take a longitudinal view of the data.

BI quadrants

As we determine which BI projects to take on, it will be critical to understand what we’re trying to achieve, and the size of the investment which will be required to meet that need.  If we limit ourselves to the easy projects in the lower left quadrant of the diagram, we can deliver lots of incremental value, but may miss out on understanding either how data relate to each other or how data is changing over time.

On the other hand, if we focus only on projects that require snapshots of integrated data, we need to be aware of the sizeable investment in time that will be required to be successful.

What will be key is to develop a balanced portfolio of projects that identifies the quick wins in the lower left quadrant and lets us reflect on the opportunity cost of the larger projects–finding a balance.

If you’re interesting in seeing how Tableau is being used right now, take a look here:

Undergraduate Admissions Reports (these views required snapshots of data from LUSI to show how undergraduate admissions progresses compared with previous years).

Current Research Projects (unfortunately, this will be blank unless you are a principle investigator)

Research Grant Income (a scatter chart comparing various universities)

Thinking about Careers (showing how mature people’s thinking is about their careers)

These visualisations just scratch the surface of what we could do with data visualisation.  As next steps we should:

  • Think about how best to catalog visualisations as they become available (to help foster and maximize the benefit of continued organic growth of data visualisation)
  • Determine what (if anything) we need with respect to a policy about who should be able to see what
  • Brainstorm, with a wide group of stakeholders:
    • What visualisations they want
    • How they’d benefit from those visualisations
    • How those visualisations might change our business processes
    • The relative size of the investments required to deliver those visualisations
  • Build a high level plan for the delivery of those visualisations (in descending priority order) and the related changes to business processes
  • Determine what resource needs there are (if any) to deliver those benefits at an acceptable rate