Local Barriers to Data Analytics for Teaching
By Dr Chris P. Jobling
The biggest barrier to exploiting data analytics for teaching and learning at my institution is lack of access to the actual data. The useful stuff that we are required to analyse and reflect on is packaged for us either in PDF reports or as unstructured tables on web pages. Presumably this is done to make our analysis easier, but it also has the effect of aggregating and filtering the data into forms that the University finds useful, or assumes that we will find useful, or perhaps are most useful for national statistics agencies. Unfortunately this makes drilling into the data difficult and limits the opportunities to gain new insights that access to the raw data would afford.
As a small experiment, I tried to copy some data, presented to me on the University’s Intranet as an HTML page containing numerous tables, and paste that general statistical information into an excel spreadsheet. I then tried to load that data into (Google) OpenRefine and Tableau. Both tools take the first table as their basis and a great deal of additional work would be required to get this into a form that would allow the analysis features to come into play.
This convinces me that the first stage of the data analysis cycle, getting the data into a structured format that can be used for analysis, will be a considerable challenge.
I know that the information comes out of a students record system (essentially a relational database) because one of the statistics pages documents the SQL queries that generate the data. Why this information cannot be sourced as a CSV endpoint (as some data such as class enrolments are) is an interesting question that I intend to ask.
On a positive note, senior management has heard of data analytics and is keen to exploit it to improve student outcomes. The problem is that because the data is in closely guarded silos distributed across the institution, they are vulnerable to the persuasive patter of the smart salesman who comes a-calling telling us that are problems are so unique that the only way forward is an expensive data analysis tool or a customised plugin in for the managed learning environment.
“putting powerful tools into the hands of individuals who might not quite understand them yet” (@psychemedia) is probably something that only the most enlightened university manager would dare to do.