When you can measure what you are speaking about, and express it in numbers, you know something about it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely, in your thoughts advanced to the stage of science.
Lord Kelvin
Introduction
The above quote from Lord Kelvin sets the tone for this blog post. This is also true about the projects and programmes I’ve worked on: how do you measure success?
The Problem (To Solve)
Many years ago, I remember visiting a colleague in her office. She was of European origin with a charming demeanour and strong (and often very correct) opinions on online teaching. Needless to say we got on extremely well.
I remember taking a peek at her screen, on one was our Learning Management System, the other a word document with a check list. It looked like some form of manual audit, trawling through each course and ticking whether certain resources existed.
My initial reaction was that of surprise: why isn’t this automated? Are we still in the 90s? Why go through hundreds of courses manually to check if they have resources? It’s time consuming and tedious. I remember telling my colleague: I’m going to find a way to automate this whole process.
The Solution
A few months later we gained access to a duplicate of the Learning Management System internal database (the data dump was hundreds of gigabytes in size). In addition it was not designed for reporting, it is one that’s designed to be the core or engine of a live system.
The data inside it was nearly incomprehensible. So I went to work in analysing it and designing queries: a great excuse to dust off my SQL skill set. I designed dozens of queries to extract useful data from the system. By using inner joins and a lot of trial and error I was able to design queries that extracted useful information from this database.
Of particular relevance was a query that scanned every single course that had existed in the system (20,000 plus at the time) and give a numerical breakdown of every resource type. With a single query I was able to extract the information that my colleague needed: something that was manual and took her weeks to do.
Over time I refined the query to add the course name and URL to the tabular data. The query was then saved as a database view.
I provided the data to colleagues in the faculty and they were able to use the data to produce basic PowerBI reports, but more importantly identify online courses that required a little ‘extra help’ and offer it to academics. It was immensely successful and an example of a great data driven decision.
PowerBI and Database Connection Magic
At the same time, I was using PowerBI. What if we connected PowerBI to the database with the ‘view’ and drew the data into the reporting software? I can then group the teaching tools into categories and create visualisations to understand the data better.
After several iterations I created a report that I was able to present it at a University roadshow. My pet project had evolved into a full blown enterprise project. Initially they called it ‘LEO audit’ but I took issue with the name, I suggested: ‘Using LEO’, which is far more ‘academic friendly’.
During one of the roadshows I presented the report on the PowerBI app on an iPad, a senior manager raised his hand and asked: ‘Can I have this now, please?’. If this wasn’t confirmation for the demand for data I don’t know what is.
Below is a sample of the report I created, again for an optimal experience please use a laptop or desktop to view the report (use the full screen option).