A first look at the UK HE longitudinal education outcomes dataset with Microsoft Power BI

13/07/18

How easy it to compare the earnings of graduates from different universities? Does a computer science graduate from Durham University make more money than one from the the University of Sheffield?

The Department for Education’s longitudinal education outcomes dataset is a good place to start (data available here). It’s a quick job to convert the data into an Excel table, fix the formatting on some of the columns (whole numbers rather than text) and then import it into Microsoft Power BI Desktop.

I went for the simplest report - just universities, subjects, years after graduation and earnings. There is more data in the source spreadsheet but at this point I just want to get my head around the data and create a very simple report.

This interactive report below is the end result. One of the great things about Power BI is that you can publish to the web. You then get a bit of html that you can embed in your website or blog. (hint - click the double arrow at bottom right of the report to go full screen)




It’s interesting to select a subject and then compare graduate at different universities but almost immediately you begin to spot issues and ask further questions.

The subject areas are broad - based on the Joint Academic Coding System (JACS). The notes that come with the download flag this up - within each subject area there is a diverse set of subjects. So you’re not really comparing degrees.

Another thing I wondered about when I compared say computer science at Durham University and the University of Derby (big difference in top end earnings after 5 years) was effect of where the graduates are employed. I know Durham pretty well and I bet most of its graduates move away (London?). Does the University of Derby have a higher proportion of home based students who stay in Derby after graduation? That could also make a difference to earnings.

There are some other columns in the source data that give more context. Next time I’ll look at how you could create a slightly more sophisticated report that compares the earning of graduates from similar universities - universities with broadly similar proportions of disadvantaged students and students similar prior attainment.

Mike