Learning Analytics is about leveraging a framework for systematically collecting, processing, enriching, analyzing and visualizing data generated by a learning ecosystem to address concrete business learning issues.
It is important to keep this definition in mind because it emphasizes that Learning Analytics is only a tool, a decision support service.
It therefore involves identifying business issues and specific questions about its learners before embarking on data visualization. Moreover, it is often necessary to take a step back from the interpretation of these data in order to avoid too hasty correlations.
Indeed, Learning Analytics encompasses not only a technical framework but also a whole system of thinking around the data, knowledge specific to the organization and its learning environment. All of this is needed to fulfill the fundamental purpose of Learning Analytics, which is to improve the learning ecosystem of an organization.
In this article, we will give you a solid overview of the discipline and some tips to get you started.
Before we can get into Learning Analytics itself, we need to talk a little bit about the data we generate in e-learning and how the implementation of the xAPI standard brings new possibilities.
In short, xAPI data is stored as “User + Verb + Object” in what is called an LRS, a kind of giant database of a company’s learning data.
This format is particularly interesting because it allows, above all, to have a common nomenclature (syntax) for your entire learning ecosystem. Indeed, who hasn’t already had the problem of trying to consolidate reports from different platforms (languages, your LMSs…) where in one case the training is “completed” in the other “completed”, in short a puzzle?
Although it is possible to do Learning Analytics without having implemented the xAPI standard and a LRS, this implementation improves so much the quantity (capture of informal learning, keywords typed in a search…) and quality of the data for a moderate cost that it would be a pity not to be interested in it.
Processing and enrichment
The LRS database basically answers the question “Who does What?
These xAPI data can therefore be supplemented by additional databases that can help to deepen the analysis. In particular, other databases can help us learn more about the “Who” and the “What”.
For example, if your solution includes user profiles, you can connect it to your employee directory. Or have a database linking digital learning objects to skills. We then talk about HR data lake, projects that we see emerging internally in several large groups and managed by IT, but which are often difficult to start up because internal IT contacts do not have the knowledge about xAPI and traditional LMS editors do not offer this type of service to their customers.
For the more advanced ones, depending on your ecosystem, you can incorporate user clickstream into your applications, their geolocation data, or even more interestingly cross-reference this learning data with business data (from a CRM or other). Thus it is possible to have the correlation between trained population and increase in turnover for example.
In our current research work at Bealink, we are trying to go even further with regard to digital content (articles, videos, e-learning content, etc.) using Natural Language Processing techniques to automate the generation of the metadata required to enrich the data in the training catalogue.
In short, you have already understood this, and at the risk of repeating ourselves on this stage of processing and enriching digital learning data, it can only be done outside an LMS-type platform.
Now that we see more clearly in our data and our various data sources, let’s see what we can get out of it.
We can analyze the data collected, processed and enriched at several levels.
Especially for descriptive analysis, a good visualization is essential.
A packaged Learning Analytics solution will provide you with its own visualization. There are a few on the market, but the problem is that these solutions are often linked to the LRS that goes with it (they sell both LRS + data visualization at once) and it is very complicated to cross-reference business data for example. A data scientist, a consultant, or our teams can help you create your own visualization models on your own LRS thanks to a tool we have developed and that we make available for free to all the learning aficionados that you are.
Example of a visualization of the connection data of one of our clients, we can clearly see the peak usage (black) of digital solutions in period of COVID. The visualization is under Power BI (a Microsoft tool) often integrated in your Office 365 suites.
You now know more about Learning Analytics; and you will surely realize the potential it has to continuously and automatically improve your learning ecosystem.
To take action, we strongly recommend, if you don’t already have one, to have an LRS in your learning technology stack.
Then, to get started and if you do not have any special requirements, you can turn to packaged solutions for analytics.
For more personalized and efficient solutions, you can turn to a data scientist, your internal IT, a consulting company or our company to help you build the famous analytical framework we talk about throughout this article.
Then, it is about your knowledge of your companies in order to determine the right questions to ask yourself about your learning strategy because you never learn just for the sake of learning, there is always a goal behind it (compliance, employability, salary increase, competitive advantage, internal mobility, innovation, talent retention, productivity, turnover increase…). But for once, you will start to get the tools that will legitimize your discourse on the importance of maintaining an ambitious learning strategy.