The Open Education API is a standard for sharing educational data. The OOAPI allows data such as grades and schedules to easily be opened using a tool such as a student app. I received the request from the Learning Analytics project, part of the innovation programme Customised Education, to explore how the Open Education API can be expanded with learning analytics data. This would allow data regarding study progress to become part of an app.
Tom Kuipers: 5 ways the OEAPI can be used in learning analytics.
1. Using the message structure of xAPI
I researched the xAPI message structure. xAPI is a pre-existing protocol for learning analytics. xAPI statements are short, structured messages about learning actions such as ‘student logs in’ or ‘student clicks on a link to a reader’. All of these individual statements are collected in a Learning Record Store.
The question posed by SURFnet was: can we process it in the OEAPI as well? What will we encounter? My conclusion was that the xAPI is well made. My advice: it is better to use the protocol than to invent something entirely new yourself. You can see more technical information about my project here.
2. OEAPI as an access port to learning analytics data
The OEAPI acts as an access port for various other systems such as a grade system or a schedule system. It can also serve as an access port for requesting the learning analytics section of a Learning Record Store. The OEAPI is the central point of entry. We have created a test setup for you to see how this mechanism could work. Statements that you would normally request in the Learning Record Store can be requested in the OEAPI – you receive the same kind of response. You can run this demo environment using dummy data on your laptop without needing to actually install anything yourself. The demo is available on GitHub.
3. Recipes in OEAPI
You need recipes for the statements, i.e. agreements about the structure of the messages. It is important that everyone uses the same verbs or terms. If one person says: ‘student clicks on link’ and another says: ‘student follows link’, it has a negative impact on communication. The Open Universiteit Nederland has offered suggestions for recipes for frequently carried out activities, such as taking part in a MOOC or logging in to a digital learning environment (click for more information).
The OEAPI can play a role in ensuring that all faculties use the same terminology. You can give the OEAPI the task: give me a list of the verbs or recipes used at this institution. Consistent use of recipes within an institution or among different institutions increases the value of the data and the analyses based on this data.
4. Adding a business layer
The focus of this assignment is on collecting learning activities. However, the data that you retrieve from the OEAPI will be of little meaning to the end user. Only after analysing the activities can the student view their progress, for example in a dashboard or as a traffic light in an app. At this point, the OEAPI primarily contains data, but has no connections between the data. Analyses of the data from the Learning Record Store take place in a Learning Analytics Processor, with which you can draw conclusions about an individual’s study progress. I think that the OEAPI could add a business layer by offering access to analyses in a Learning Record Warehouse.
5. OEAPI as a generator of new learning activities
It may be a bit meta, but it is interesting nevertheless: the OEAPI itself is a generator of new learning activities. Whether a student requests their grades or their timetable is indicative of their level of involvement in their education. You could use the OEAPI with a learning analytics activity statement such as: Checked: marks. Checked: schedule.
Using Tom’s work, SURFnet has positioned the OEAPI in the learning analytics architecture developed by SURFnet together with various higher education institutions. SURFnet will research how the OEAPI can be used in practice. The UvA is continuing to improve the learning environment of its students and teachers by implementing learning analytics.