Tech Tuesday: Dive into Big Data


The Future of Learning Data

Historically, Learning and Development (L&D) within companies has always wanted to keep the reigns around all learning content by collecting and managing data within a single Learning Management System (LMS).  This approach has really gone the same path as the Enterprise Relationship Management (ERM) systems that were implemented in the late 90s-2000s.  Huge systems that take teams of people to install, configure and get up and running only to fall short of the expectations of the system due to limitations.  The natural next step has been to bring in a consultancy to help build out the extra 20% that the LMS did not provide to try and get the LMS to a state where it would be usable (and useful) for the entire organization.  But, with the evolution of cloud technologies, L&D has started to realize that it is ok (if not preferred) to work with smaller systems to try bite-sized chunks of learning objectives using simpler tools such as mLevel to provide a micro-learning approach.  This evolution has helped L&D in two areas: (1) quicker roll outs and feedback and (2) new ways to gather data around what employees are proficient at and may need additional help.

mLevel has evolved the data analytics and reporting platform from simple data exports, to colorful dashboards that provide data about learning objectives at the course (mission) level, to more advanced integrations into larger data systems that handle the major learning protocols including LTI and xAPI.  During this evolution, the amount of data has changed as well.  In the beginning most L&D staff only cared about whether or not an employee completed a course or not, now the staff is trying to provide additional value by being able to truly understand what each person knows and doesn’t know at the most granular level.  Enter XAPI and the LRS. These newer technologies allow people at all levels of a company to understand how well their employees understand their business by way of very detailed reports that can be sliced and diced in any preferred way.  The data that is generated for the LRS through the XAPI protocol will help L&D have a view into what a learner interacted with, how the interaction went, and some color (e.g. added metadata) to help the team gain insight into how to evolve the learning program in the future.

mLevel and the Data

With mLevel, the system tracks events while a learner progresses through each and every learning activity in the system.  Events such as when a question is displayed, which answers are displayed, which answers are selected and even when a learner pauses a game are collected and stored within the mLevel platform.  All of this data is queued up during the interaction and sent as a simple JSON structure for later parsing on the server for consumption by learning administrators.  Even with all of this data being available in the mLevel system, the common requests seem to come down to a simple view of whether or not a learner has Completed a mission and when it occurred.  There are two ways for the mLevel platform to provide that information: (1) the chart below and (2) a data export specifically centered around these simple data points.


While these methods do provide the answer to the basic “Completed or not” question, it generally means throwing away all of the extra data that is being collected during a learner’s interaction with the learning activity.  Sure, mLevel does provide some data around which questions and topics are understood collectively (across the population of users), but it does not directly handle showing how well individual users have grasped the topics related to the learning program.  Once again, enter XAPI.  mLevel has hooked up an event engine to create XAPI statements to provided details around specific interactions with the learning activity.  XAPI statements are simply JSON structures of unstructured data to outline exactly how a learner interacted with the platform.  By using XAPI and Microsoft Azure Event Hubs, mLevel has built out a system that can be consumed by any LRS on the market by simply setting configurations within the admin console.

mLevel Platform Event Data Flow

As depicted by the flow below, the data within the mLevel platform is persisted within a relational data store, but also sent into an Event Hub for consumption by any subscribed listeners to the various data events raised by mLevel’s Game API.  In the upcoming series, we will walk through how the Game API and event engine processor the JSON inputs to create all of the data necessary for systems like an LRS to process XAPI into meaningful reports.


Come back next week, to see what data looks like coming into the system and what it looks like when it is ready for consumption by XAPI consumers.  In the weeks following, we will walk through how to setup an internal Learning Locker instance to start to understand how to begin taking the next steps to big data for learning activity data.