E-learning and data: two peas in a pod
Download as PPTX, PDF0 likes82 views
The document discusses learning analytics, which is defined as the measurement, collection, analysis, and reporting of data about learners and learning environments to understand and optimize the learning process. It outlines why learning analytics is important to comprehend future work needs, make training behaviors and needs more visible, and inform learning. The document then discusses how to implement learning analytics through gathering relevant data points, looking for data sources, starting small, and measuring the right things. Finally, it outlines a future for learning analytics that includes diverse learning times and places, personalized learning, free learner choice, and more humanized learning.
E-learning and data: two peas in a pod
- 2. Wim Van Borm Senior learning & development specialist
- 3. LEARNING & DATA two peas in a pod
- 4. AGENDA WHAT is Learning Analytics? WHY Learning Analytics? HOW to do it? Into the FUTURE!
- 5. WHAT is Learning Analytics?
- 6. Hailed as the next BIG thing? Solve all the L&D problems? But What is it! BUZZ and HIPE
- 7. DATA and it’s BIG All over the place Impacts our day to day live
- 8. DATA and it’s BIG All over the place Impacts our day to day live
- 9. DATA and it’s BIG Traffic camera’s and systems • Pattern recognition • Adjusting the Flow
- 10. DATA and it’s BIG All over the place Impacts our day to day live
- 11. DATA and it’s BIG Banking Solutions Financial management @ our fingertips • Smart ATM’s • Virtual assistance • Chatbots
- 12. What is BIG data – The information Continuum Cartoon by David Somerville
- 13. BIG data • Scale of data • 90% of todays data – created in the last two years • Every day creation of 2,5 quintillion bytes = 10 million Blue rays disks • 40 trillion gigabytes of data by 2020
- 14. BIG data • Forms of data very divers and continuously increasing
- 15. BIG data
- 16. BIG data • Analysis of data flow • frequency of incoming data
- 17. BIG data • Uncertainty of data is the data trustworthy • Origin • Authenticity • Trustworthiness • Completeness • Integrity
- 18. Big Data in Learning Big DATA finds its way into Learning • Through IOT • Through LMS • Through Xapi • Through performance measurement • … Automatically collected A system is in place that automatically extracts and stores the relevant data that is generated 3 Continuously analyzed Information is relevant to human well being, development and can be analyzed in real time 4 Passively produced Data is by-product of integrations with digital services 2 Digitally generated Data is created digitally, not digitized manually and can be manipulated by computers 1
- 19. WHAT is Learning Analytics • The Measurement • The Collection • The Analysis • The Reporting of data • About: • learners • their contexts • for purposes of • understanding • optimizing learning • optimizing the learning environments • Achieving business goals
- 21. USING DATA to INFORM LEARNING and WORK
- 22. Comprehend the future of work Change is all over the place • New behavior • Adjusted behavior • Additional behaviors • … Behavior = knowledge + application We need to know who does need what. One size fits all is outdated!
- 23. Make behavior visible Visualize training behavior. • People not completing courses • People who click trough the learning in 5 sec. • People not being engaged in the learning • …
- 24. Make needs visible Visualize training needs: • Where are people today • What are the needs of the organization • What are the needs of the individual • What is the gap between company needs and capability of the individual? • …
- 26. Different levels Corporate Centric Analytics • General perspective of the organization • Provide framework • Organization effects People Centric Analytics • Helps us to personalize • Focus on the individual • Correlated to the corporate needs
- 27. People Centric Analytics Requires: • Different data type • Different approach BUT Average data quality and availability is low
- 28. Gathering data points Know what data you need and why you need it! - Understand the problem - Understand the need - Understand the goal SMART
- 29. Looking for data Knowing where you can find this data Create data entry points
- 32. Training evaluations • Success case method • Learning Transfer Evaluation (L-TEM) • Trainer’s balanced scorecard • …
- 33. Start Small Taking a small but relevant sample
- 34. Gathering data points Create a sample report Request feedback from stakeholders Does it work for them Refine step by step
- 35. Is it worth measuring Possible to measure all ? Is it useful to measure all ? Where should our focus then be?
- 36. Is it worth measuring URGENT NOT URGENT NOTIMPORTANTIMPORTANT QUADRANT 1 Urgent and important DO QUADRANT 2 Not Urgent and important PLAN QUADRANT 3 Urgent and not important DELEGATE QUADRANT 4 Not Urgent and Not Important ELIMINATE
- 37. Is it worth measuring URGENT NOT URGENT NOTIMPORTANTIMPORTANT QUADRANT 1 Urgent and important DO QUADRANT 2 Not Urgent and important PLAN QUADRANT 3 Urgent and not important DELEGATE QUADRANT 4 Not Urgent and Not Important ELIMINATE URGENT NOT URGENT NOTIMPORTANTIMPORTANT QUADRANT 1 Urgent and important DO QUADRANT 2 Not Urgent and important PLAN QUADRANT 3 Urgent and not important DELEGATE QUADRANT 4 Not Urgent and Not Important ELIMINATE
- 38. Measuring is Not an exact number Specific point(s) in time Representation of the reality
- 39. Measuring is A KEY ingredient in sound decision making Reduction in uncertainty based on observation • EDUCATED guessing
- 40. Rule of 3 Estimate the chance of something that has not happened yet with a 95% probability.
- 41. Predicting ranged What is the: • Minimum • Optimum • Maximum
- 42. Predicting ranged Project on the table • 400 courses to be updated /year • How many should we update • Which courses should we update • In order to support the user base in a proper way.
- 43. Predicting ranged Project on the table • 400 courses to be updated /year • How many should we update • Which courses should we update • In order to support the user base in a proper way.
- 44. Into the FUTURE
- 45. Diverse time and place Learn @ different times Learn @ different places Remote self-paced learning
- 46. Personalized learning • Tools adapting to the capability of the learner • Allow practicing for those that need it until they reach the required mastery • Specialized reinforcements during the individual learning processes.
- 47. Free Choice • The road leading towards the destination should vary per learner • Learners should be able to modify their learning process • Different tools, different programs, different techniques • Based on their own preference
- 48. Humanized learning increases the relevance of content to the learners and improves one’s motivation to login week after week.
- 49. ?
Editor's Notes
- #8: Music streaming These services collect crowdsourced data from user feedback and behavior. They curate through algorithms playlists for each individual user Spotify offers each user a personalized playlist.
- #9: NETFLIX Levering customer data to define the most popular genres, concepts etc. Applies predictive modelling Netflix provides a rating for each title based on Netflix thinking you’ll like it or not.
- #10: Traffic camera’s capture data and help the improve the traffic flow in cities. Traffic patterns recognition Act upon the patterns or changes in the pattern Help us to create a better traffic flow.
- #11: Big data gets you where you need to be on time. Numerous real time data points and metrics go into measuring the traffic Apps use cell location data to determine the fastest route to your destination + provides estimated time of arrival. On top of that the calendar app will notify when you should leave to be on time for your appointment based on real time data.
- #12: Big data in banking solutions Financial management @ our finger tips We don’t have to go to the bank any more to get our stuff done.
- #13: Data is generated by numerous people Together they will provide you information This information will provide you knowledge From which you can derive insights and finally wisdom. Look @ Waze collects the position of the cars and the inputs provided by users + a massive other amount of data sources This data provides information - I’m here - GPS location Which will provide us with the knowing of the information – by chairing information we know. Combining all this knowledge and analyze it we might see patterns and other. This means we will create insight based on the knowledge Connecting the dots Google is all well and good, but if you’re using the wrong search terms, you’re not going to find what you’re looking for no matter how many times you click the Search button. You have to imply a certain amount of insight and use some guesswork (“I wonder what most people would call that?”) to get what you’re looking for. Google can’t do that thinking for you. The movie “limitless” Our Hero, Eddie, is a smart but unmotivated fiction writer. He takes a drug, and suddenly he’s brilliant. But not because he’s suddenly taking in a huge new ocean of data; no, it’s because he can make connections between data that was already there in his noggin, and create new ideas and new conclusions from it. Such is also the brilliance of Sherlock Holmes, Mr. Spock, Herocule Poirot, and many other “genius” fictional characters. It’s not the data, folks, it’s the connections and conclusions and deductions. And eventually we might learn something from this and achieve wisdom;
- #14: Big data analytics is the process of examining large data sets to uncover hidden patterns, unknown correlations, market trends, customer preferences and other useful business information. Scale of data every day we create 2,5 quintillion bytes of data = 10 million Blue ray disks
- #15: Tremendous amount of variety of data – significantly increased and keeps increasing continuously
- #16: Tremendous amount of variety of data – significantly increased and keeps increasing continuously
- #17: frequency of incoming data significantly increased
- #20: Learning analytics is the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimizing learning and the environments in which it occurs. This in order to support the business goals.
- #29: Knowing what you need – what you are trying to resolve or what goal you are trying to accomplish will help you to define the data and the metrics. Asking questions does not really cost serious money. But by asking questions and analyzing the results we can draw some conclusions. These might cost really money but then we know why it is needed So this helps to get a budget- Be specific in what you are aiming for. SMART goals M Measurable
- #30: Sometimes it’s hard to get the data. We don’t have access to it Sales does not want to provide it. Or I can’t work with the tools to give me the data Network within your organization. Talk to people who might help you with this and yes buy them lunch to talk to them. But also start using feedback happy sheets But also using more in depth methods - Kirkpatrick Evaluation Model
- #31: But also using more in depth methods - Kirkpatrick Evaluation Model 4 levels: Reaction – after the training occurred Learning Behavior Results - impact
- #32: 5 levels
- #33: They all provide datapoints from a learning perspective.
- #34: You do not from day one need all the data. You can start with a small but relevant and applicable dataset which is representative for the entire dataset. Based on this make a sample report and present this to the stakeholders involved. This quick sample will provide us with good insights and will help us to improve / adjust
- #35: Various stakeholders – different requirements but also different responses to the provided report! Match your report and data with the audience requirements. Try so speak their language. Not all stakeholders will respond in a positive way to your report. Some people do not want to listen to the data Some people will listen to the data but they have already made up their mind Some people will go way beyond the requested. Some people are emotionally attached to their program/project Make sure to engage the receivers but equally the requestors for the data. Align upfront with the various stakeholders
- #37: Which of the 4 quadrants of the change management matrix would be applicable for learning analytics?
- #38: Things that are not (yet) urgent but very important. Will allow us to gather the data – and make these visible prior for them becoming urgent and thus allowing us to act prior
- #41: Helps you to set the requirements for the data sample size. The larger the sample size the smaller the chance on an error in a random course will be if no errors were found in the sample size. 30 courses reviewed no typo’s found then by the rule of 3 we can state that there is a 10% change of typo’s within a course created. 300 courses reviewed no types by rule of 3 chance for a typo in a course is 1 %
- #42: What is the minimum / the max ==> but we ail for the OPTIMUM Project of the table: not being possible to update all 400 courses per year. How many and which courses should we update in order to support the user base in a proper way and provide marketing with enough new stuff to run a valid campaign. What is the minimum What is the max What is the optimal
- #43: Project of the table: not being possible to update all 400 courses per year. How many and which courses should we update in order to support the user base in a proper way and provide marketing with enough new stuff to run a valid campaign. What is the minimum What is the max What is the optimal
- #44: 400 courses are spread out over 40 categories 10 course per category What are the high value courses What are the low value courses ==> measurement on how many students bought the course/ attended the course ==> measurement on investment into the course (creation - marketing etc) ==> measurement on the live cycle of the content Short shelve life Long Shelve life ==> the possible audience ==What is the potential audience size Course types ==> measurement is the course a compliance and does the content vary fast or not / yearly not yearly Doing 40 courses per year is perfectly achievable. Various datapoints are required in the above example. Even for determining the minimum en and max and the optimum.