Your Source for Learning
Technology, Strategy, and News
ARTICLES       RSS feed RSS feed


email print

The Experience API (xAPI): A GPS for Learning

by Michael Hruska

April 25, 2016


by Michael Hruska

April 25, 2016

“We are about to change the game through interoperability of data about human experience as the xAPI expands in the world. We need to keep finding ways to work together to bring the GPS for learning into the world.”

Learning has changed. Today, our interconnected lives are filled with technologies and new ways of connecting to information and one another. This shift, created by mobile and social along with nearly ubiquitous connectivity, has evolved the world as we know it—and, more importantly, it has given birth to new capabilities for the learning ecosystem.

Today’s learning ecosystem is rapidly growing, extremely complex, and overflowing with potential data. Individual employees can access multiple systems on and off the job for formal and informal learning—anything from LMSs to MOOCs to virtual worlds to simulations to classrooms to mobile devices to YouTube. Learning can happen nearly anywhere at any time.

Yet, while this is possible, learning the right thing at the right time is the real challenge. Simply having access to so many systems and experiences creates noise without creating value. Knowing exactly what someone should do next is the core of what we should be most concerned with. The key is stacking experiences in the best order to get us to our goal.

Affordances of mobile, social, and emerging specifications like the Experience API (and others) are positioning a revolutionary learning ecosystem that offers individuals help in navigating all those possibilities. We are now on the cusp of realizing the value and potential of this development.

GPS applications offer clues to a navigational model for learning

When the creators of the Global Positioning System (GPS) navigation system were designing it, did they anticipate that businesses would emerge from the ecosystem that GPS made possible?

For example, Uber relies on the affordances of the navigational ecosystem—a place where people travel as nodes tracked by sensors within maps supported by guidance systems. But Uber is doing much more. It is building atop the sensors and flows of data at new levels in order to understand the patterns in a different way. This understanding allows the company to connect all passengers with drivers to get them to their destinations. And if there aren’t enough drivers, Uber finds ways to push the pool to meet the needs of the riders—Uber’s understanding of patterns creates a balance within the ecosystem. This balance is causing new value to emerge from the network. In fact, there are already many business models that take advantage of the affordances of the GPS ecosystem, and the number of examples is growing. 

Navigation using GPS relies primarily upon the ability to define a destination. What makes GPS navigation really smart is that it can also include:

  • Maps of the terrain
  • Data from other travelers before you
  • Data from other travelers right now

These data, in concert, allow some amazing things to work automagically.

What does this have to do with the xAPI and the learning ecosystem?

Are there value creators that can emerge from learning ecosystems, similar to the way Uber emerged from the GPS ecosystem?

What’s missing is one step further toward a path. A path is a series of steps to achieve a goal. You could even say that what is needed is a GPS for learning. The real and defined need in education, learning, and training, across the education and corporate landscape, is a chain of steps to get you to a destination.

Steps are just sequences of locations and directions. Directions focus on destinations. In the classic case of higher education, for example, destinations are typically degrees. In the case of occupations, destinations may include competencies. In business, destinations may be accomplishments that are the result of applied competencies and experiences.

We also need a map, and we need information (data) from those who have sought the same destinations. But there’s a problem.

We don’t have the map (yet)

GPS works because sensors are working in concert to produce an understanding of position. Because of maps, we are able to define the steps between where we are and where we need to be—and our navigation system can help us get there. We still have to work for it by driving, and most importantly, we need to stay on the road, follow the rules, and not run into anything along the way.

What are our maps for learning? We don’t really yet have a “global map.” We have a lot of data, but we don’t have a global standard. Different products or projects try to tackle this at different levels. Efforts like LinkedIn Skills lists, the Learning Registry and even Wikipedia are trying to map the world’s knowledge and learning. We don’t yet have the whole map—we essentially have only local versions available within a company or domain. Competencies and experiences (content) are the key to making the map. Think of competencies as intersections of pathways or roads. They are the little destinations that someone can get to from a number of directions while the content provides the steps between the intersections.

Competency models and maps take many forms and have no unified data models in practice. Currently, we have the equivalent of hand-drawn maps of villages or towns. We have “cave-like” line drawings of steps as well, such as in onboarding. We need better ways to describe and map competencies (yes, and knowledge, skills, behaviors, and measures) and their relationship to content or experiences.

Projects like the Learning Resource Metadata Initiative (LRMI) and numerous previous attempts at mapping digital objects and metadata (Handle System, Dublin Core, IMS Learning Resource Metadata, and others) aim to attack the mapping problem from different angles. The Simulation Interoperability Standards Organization (SISO) is also standardizing further work on modeling human performance using Human Performance Markup Language (HPML).

Although work has been done, and continues to be done, a global model isn’t enabled without a real way to describe the intersections and roads in one common language. The Advanced Distributed Learning (ADL) Initiative believes there might be a way to standardize and associate the maps. ADL’s new project, the Competency and Skills System (CASS), aims to provide a larger framework for associating competencies.

These are great moves forward, but we still need interoperability of this data to connect intersections and steps to make a real global map. Without the real map, we will just have relative positioning. Allow me to offer another set of examples that may give you a better picture of what is involved.

Destination ISS (the International Space Station)

Depending upon whose definition you use, about 550 people have visited space; 222 of them have visited the International Space Station. As of today (April 25, 2016), there are about six people in space. Dreaming about getting to space and actually getting there are two different things. There are many common destinations in the paths of each of the people who have been to space, but there are also many different paths. Wide varieties of military and academic backgrounds in different countries all led to a set of core experiences to allow this group to do what few have done.

In the same sense, there are pathways to travel through education and experience, and many intersections and many turns exist. Though people start in different places and take different roads, they eventually end up at the same intersections and same steps.

Setting a destination is important to any journey. If learning is going to be like a journey, we need to be very specific at defining destinations. If we have a map and destinations, getting there will be quite a bit easier. The path to get there will be clearer, and knowing where we can go from any point will be much more obvious. But first, we have to know where we are.

Position: Where am I?

Of greatest importance to GPS is the latitude and longitude grid. This grid tells us where we are (and also where we are not). The corollary in learning is the profile. The profile includes a history.

With the xAPI, we have a way to describe the different waypoints (landmarks) in a person’s history, places where they have been (experiences), and intersections or destinations (competencies) they have passed through. xAPI data can capture the step you are taking (content) and also information such as your performance (context) toward an intersection (competency) to get to a destination (course, badge, degree, job), even if that’s a job on the space station.

Schools provide destinations

What are schools? They are groups of people who have a “recipe” that they have agreed upon and that has created outcomes or destinations that everyone agrees are generally good. And the people who have experienced the outcomes then become evangelists for the outcomes or the pathways.

A school is a set of learning experiences that have a high probability of achieving a desired result. That is, these experiences constitute a collection of curated and potentially productive pathways toward a destination.

Think of a school as a map to get from where you are to a common destination where others have already gone. Most importantly, it is a place from which one can move to the next destination with a high degree of probability.

Now, there is a deeper opportunity.

There are many destinations in the world, such as being a doctor, being an artist, being a biomedical engineer in Austin, Texas, or being like my friend Pat. More importantly, questions such as “How can I be like a certain co-worker who has X, Y, and Z skills that I don’t have?” are quite interesting. Most important is knowing what you don’t know that can get you to your destination. It’s really all about skills acquisition. Skills come from knowledge and practice to build ability. These all come from the right sequence of steps or experiences toward the destination.

What are the destinations that we define in life? Where do we store them? The answer to both questions is “many places,” and that brings us to the important problem of data interoperability.

Data interoperability across the ecosystem

New efforts are emerging around data interoperability in this future ecosystem. Aaron Silvers and Megan Bowe of MakingBetter established the Data Interoperability Standards Consortium (DISC) in October 2015. They are working toward something bigger, building out from the xAPI in a way that will cut across industries and beyond L&D.

“Since 2014, there’s been an industry effort to move xAPI forward as an industry. … The spec needs an organization that addresses the bigger world xAPI is part of,” says Aaron. He is referring to the fact that maps or positions are just part of a much larger opportunity to create the GPS for learning. xAPI is one component, but we need more.

Regarding building the future of GPS-like ecosystems, Aaron says, “With data interoperability we’re looking to enable a data environment where we can successfully design, build, and grow systems on common expectations of data. This enables increased interoperability, data ownership, and advancements in analytics.” DISC is focusing on three main groups: tool providers, professionals who build and design systems, and individual data owners.

Aaron adds, “The focus for data interoperability is larger than xAPI, but obviously xAPI is a known entity with plenty going for it and a lot for all of us to work on, so our goals for now are to make working with xAPI easier for stakeholders, developers, designers, data analysts … everyone.” DISC will be focusing on software certification in 2016, moving on to a certification program for professionals who develop learning record stores, and then focusing at another level on additional professional certification programs.

GPS and experiences

Explicating the destination is important. Acquiring knowledge, skills, and abilities on the way toward the destination comes through formal, informal, and experiential learning.

Thirty years ago, we had “great” paper maps. Fifteen years ago, traveling was a bit easier as you’d print directions from MapQuest or other sources. This allowed you to get in the car, drive, and use the specific steps and guidance directions and more granular maps. If you got off course you’d have problems, and depending upon the town, potentially real problems! Today, the world is quite different because of mobile and persistent connectivity that provides constant adjustment of position and course. With many other nodes in the system, we have far better maps and can predict success or failure on any given route using data from many nodes simultaneously.

Why isn’t this functional evolution more apparent in our learning systems?

Learning systems can and should be a lot more like GPS. They need to know where I am. They should allow me to set a destination. I should have an idea of the path or route. Based upon my preferences or other choices, I should be able to affect the route. When I’m ready to move ahead, based upon the map and other data from actual “travelers,” it should get me there via the best way possible.

Ecosystems and GPS

With the explosion in technologies and instant access to information, the opportunities for learning and the availability of learning data are massive. Systems acting in concert could be like an ecosystem to support learning. In much the same way that GPS supports our wayfinding, a system could support our guidance for learning. With all of this, we could understand the best way people have gotten there and the best way people are getting there right now, including the challenges and obstacles of the day.

Learning ecosystems are essential to capitalize on this web of data to enhance learning experiences. By placing “sensors” within learning ecosystems, we are able to create a GPS-like ecosystem. The data provided by the new technological age has the potential to propel learning and enable learning ecosystems to deliver the right learning experiences to the right people at the right time.

We are about to change the game through interoperability of data about human experience as the xAPI expands in the world. We need to keep finding ways to work together to bring the GPS for learning into the world.

From the Editor: Want more?

Watch the video of Mike's talk on learning ecosystems, given March 17, 2016 as part of the xAPI Camp held during The eLearning Guild's Learning Solutions + Eco Conference in Orlando.

Topics Covered

Appreciate this!

Login or subscribe to comment

Nice article Mike! I like the analogy of GPS. I think it also helps speak to the "diffusion of innovation" curve that xAPI is on. The kind of GPS functionality that Uber is built-on was used, refined, and championed by a hard core group of nerds - that lead to widespread adoption (and now it's a part of our daily life in a way that's nearly as pervasive as running water).

Also, I appreciated the competency modeling stuff. I wasn't aware of that.

Lastly, I think there's room for a bigger part for communities and social learning in the story you're telling. When dealing with complex and/or ambiguous paths and destinations, there's human-to-human problem solving and navigating that I think will be key.
Maps and GPS navigating a driver from her existing location to the destination of her pathway is a good metaphor for simple illustration of complex learning processes. (It is getting popular in "adaptive" learning, I also used it). Its simplicity is due to the facts that at any moment the driver can be located only at one single point of the 2D map, and existing and targeted locations are easy to define and "observe" on the map.
The real learning space is much more complex. Moreover, the current and target "locations" in the learning space are ill-observable and ill-controllable. They are not single points, but fuzzy distributed constructs, mental models. That is why another metaphor, a mental model construction (+testing+remedy) would be more practical and theoretically compliant.
I used to use the GPS and pathway metaphor for superficial illustration, but the mental model construction in my serious work for deeper learning.
The learning (and performance) space is much more complex than just GPS (in regard to goodok comment). This is the fundamental challenge and the reason that while we have come far, we still have a long way to go. Concepts like collective intelligence, readiness, or talent management can be built out so much more robustly in the future if we are aligned on design metaphors with interoperability or portability of key elements in the ecosystem. Much research on complexities like intelligent tutoring systems, knowledge skill tracing, student models, and more are key illustrations. These are much more granular and more dimensionally complicated as well. They go beyond the GPS metaphor.
The GPS metaphor is a design and functionality metaphor intended to provide a larger "map" or context. More of an overarching war cry is need to focus efforts on the future functionality and emergence of highly valuable technology.
The intersection of competencies, content and experiences is a core part of the concept, but you are 100% right that there is much deeper complexity when we get out the magnifying glass. Like any good fractal phenomenon there are scalular independent pieces of this and there are also pieces that are important at each scale. Points at one scale are maps of finer spaces at another upon zoom.
At the overarching level GPS is a metaphor, and at closer inspection other models could provide a more meaningful design or analysis metaphor.
@goodok, I 100% agree. The challenge is the paradigm shift from what constitutes L&D today to what we're starting to unlock with the more advanced uses of xAPI is a giant leap for many practitioners, managers and stakeholders who come to this site. That's why it's important to relate the concept to something most everyone can relate to, even if it is not entirely adequate.

Indeed, navigating three dimensions is a huge leap over the path making of GPS technologies, but we're talking at least navigating four-dimensions and, hard as it is to conceive, we're talking quantum and for that we have seriously inadequate metaphors to relate to.

Because stuff is hard. :)

In the learning analytics community, another metaphor for what Mike talks about at the end can be traced back a few years -- "pheromone trails" which in some ways/contexts may be more accurate, in the way that ants perform way finding based on the pheromones left by others.

Kirsty Kitto (Queensland University) discusses this as "Data Pathways."

I guess my point is that we have lots of ways to describe a similar idea, and the closer we want to get to reality, it's to use all these metaphors to triangulate the idea -- because as we learned from General Semantics, "the map is not the field." In other words, text is imprecise in describing something that is beyond text.

Anyway, thanks for giving me an opportunity to nerd out a bit here.
Related Articles