I’m often asked to talk about Education Analytics or Learning Analytics, and I use the picture above for the first part of the story every time. Not because I’m lazy, but because I think that we need to talk about the true purpose of analytics projects. We’re trying to answer questions from people inside our organisation, but often we get diverted into a conversation about ‘reports’ - what will the report you want look like? What are the fields you want on it? How often do you want us to run the report? And that conversation ends up being the focus of the work…
Creating Education Analytics stories
But we’re turning the corner - heading towards producing information rather than reams of data in difficult to decipher sheets (I felt that I needed a NAPLAN version of the Rosetta Stone the first time my daughter's report came through the letterbox at home). And if you ask my colleagues, then you’ll find that one of my biggest joys around analytics is Power BI, a series of apps and web services that return the joy to working with data. And the reason is the picture below:
It’s about the way that you create answers to questions. And although the journey has three steps, I’ve often found that analytics projects and conversations have missed out the middle one! It’s the step of talking about the story that the analysis will tell, and the building of the story. It’s forgotten because we quickly jump from Step 1 (“Hurrah, I’ve finally got the data!!!”) to Step 3 (“Let me build you the report\dashboard\scorecard”). And, in fact, I’ve often seen the cart put well before the horse by people asking me to first specify what the report will look like - weeks, months and sometimes years before they are produced.
Q&A is like searching the web for answers with a search engine - you just ask questions in plain English - but it is looking for answers within your own data. So for example, if you want to know the NAPLAN score for year 3, that’s exactly what you type…and it will take your data and display it to you.
Power BI changes the process by making it so easy to craft the data stories that instead of needing specification in advance, you either create the report together with the user, or use the Q&A feature to actually ask questions of the data with plain language questions, and explore the data to find the stories that sit within it.
Q&A in action
Here’s what using Q&A looks like with example NAPLAN data. By typing the question into the box, it looks to see if it can answer the question. In this case I typed “average score by domain”, and Q&A did a few key things invisibly:
- It looked at what I’d typed to see what it meant - in this case, it worked out that average and by were words I was using to tell it what to do with the data, and score and domain were the items of data I wanted it to work on
- It did the maths on the data to give me the average scores across all of the four domains
- It then decided how to display the data, and in this case chose a bar chart. Depending on the data types, and the volume of data, it will pick different ways of showing it. And if I don’t like it, I can simply add a phrase like “as a line graph” or “as a scatter chart” or “as a map” to change it.
If I add one more word to the phrase - below I added ‘year’ then that will look at the data again and reinterpret my question. So you can see below, it split the bar chart with scores for separate years showing:
And then finally I added the number ‘3’, which asked it to only show the answers for Year 3.
The kind of thing I’ve shown above is exactly what happen all the time when people are looking at reports - they ask to see more data, or for data to be shown in different ways, and often that means there’s a wait for a day, a week or longer whilst somebody re-writes the report. Whereas with Q&A it is simply about asking the question of the data in more detail, in real time.
Analysing NAPLAN data
Although there are lots of different ways of exploring NAPLAN data, often they produce fixed, pre-coded visualisations and reports - ie the reports that people think you should see. Whilst visualisations are great (because, to be honest, the big spreadsheets of NAPLAN data on their own are a perfect cure for my insomnia) the step beyond - to help teachers and leaders to explore the data and to feel ownership for the interpretation of it - requires something more than standard reports.
Literatu, is an Australian software company that focuses on creating a teacher-friendly formative assessment platform to collects and measures real time student learning, and we’ve been collaborating with them closely over the last few months while they have been developing NAPLAN Explorer. It’s a learning analytics system to help you to get deep insight from your assessment data, and allows you to avoid the problem of being overwhelmed with too much data and too little information.
They have created some excellent charts that are standard in the NAPLAN Explorer system, giving detailed information about growth by class, subject, skills, and progress tracking and skills gap analysis for teachers across their students and classes. The screenshot below is a great example of making complex information simple - it looks at the key skills in each NAPLAN domain, and it is simple to interpret, because the lower the achievement the larger the circle. In a nutshell: Bigger Bubbles Mean Bigger Troubles.
Use Power BI to analyse NAPLAN data
It’s a great diagnostic tool for teachers and leaders. But what happens when you want to look at the data and there isn’t a pre-defined chart? That’s where the team at Literatu have used Power BI. What they’ve added is a Power BI button into NAPLAN Explorer, and that takes you from NAPLAN Explorer straight into NAPLAN Explorer for Power BI, where you have more customised reports as well as two key new opportunities:
- Use Power BI Q&A to ask questions about your data
Similar to the example earlier, with the ability to keep digging to get more detail - so for example to be able to say ask complex questions like “average score class 3E2 ATSI” (this is the kind of thing that takes a second to write in Q&A, but could take somebody weeks to create in a report)
- Combine your NAPLAN data with other data sources
As well as NAPLAN data, schools have many other sources of formative and summative assessment data (like PAT test data), and lots of other data sources, whether it’s in the Student Information System, teacher spreadsheets or other places. By bringing your NAPLAN data in an easy-to-query format into Power BI, it makes it easier to connect it to other data sources you can bring in too
Getting Education Analytics project started with Power BI
So if you want to get started with an Education Analytics project, how would you get started with Power BI in order to learn what’s possible:
You can get started free - just register for Power BI and you can download the apps, including the desktop app (to create data sets) and the Power BI app for Windows, iOS and Android - which allows your users to access your data when and where they want, on their own devices - directly from PowerBI.com.
You can get started with the free Power BI licence and then as you get into advanced usage, switch to the Power BI Pro which is around $3 a month (and you can even get a free 30/60 day trial of that too)
For analysis of NAPLAN data, I really, really recommend starting with NAPLAN Explorer for Web and Power BI, which is a service that you subscribe to. The NAPLAN data is very complex and contains lots of detail, making it time consuming to analyse at the level teachers will want. So NAPLAN Explorer takes all of that pain away by structuring the data, creating the charts for your teachers, and then allowing you to use Q&A.
This approach with NAPLAN Explorer also allows you to deliver something with very high value very quickly to your colleagues. And then you can focus on building much more customised information with other data sources next through Power BI
Next week I’ll write about free training resources for Power BI…