 Hello, welcome back to Learning Analytics Tools course. This is the data collection part 2. In the last video, we saw data collection in classroom also in the moodle or MOOCs. In this video, we will see data collection in a Technology Enhanced Learning Environments. Tell environments is a learning environment to teach learners. It is a Technology Enhanced, why it is called Technology Enhanced. In Telly, typically compromise of learning objective that is student has to solve a complex problem or student has to learn a particular subject, a particular topic, something like that. Or you can assign a task to the student saying that in Telly, student should able to create a project or create a model or able to create a concept map, something like that. And the Telly typically have the resources to solve it, like if you are asking them to create a model, there should be a tools, simulators or reading materials, everything has to be available. Sometimes you can also have support in the form of mentor or tutor like agent or animated agents that support can be provided. In Telly, the focus is on students learning. It is not about how do you deliver the delivery content, you want to deliver using a video, the delivery mechanism or device. The more focus is on how to help the student to learn. So we can have a technology, it can act as a scaffolding. In Telly, technology can act as a scaffolding, giving feedback when the student completes level 1, the Telly can help them to move on to level 2. And when they have a trouble in level 2, they can give feedbacks to improve the learning or improve the understanding. Open ended learning environments is a student centered Telly, it is also Telly. Here the learning process is shifted from teacher to students. In Telly, the teacher is the one who is setting up the path saying that the student has to learn this task first. After completing task 1, they should do the task 2, then task 3 or they have to read this material, create this particular model and they have to go and answer some quiz questions. Here the teacher is setting up the learning path that is in Telly. However, in open ended learning environment, the learning process is shifted from teachers to students that is given a complex problem and tools to solve it. Learner explores multiple strategies to solve that problem. The teacher is not telling that you have to do this step first, you have to do the step 2 like that. We are given a problem, we will do all the tools and resources required to solve the problem. Let the students to explore his own path, you can read first or you can go and take the exam first or you can create model without reading, does not matter. It is the students exploring their own learning path. So here students sets their own goals, you know students will think about I want to solve a part of this task. Then I want to apply it and test it, then I would like to do the solve the part of second task or something. So students sets his own goals and he has sets his own planning and he might monitor his own process if the process is not working and he might re-plan all these things happen within students. So the problem is this OELE is not good for novice learners because novice learners often have difficulty to complete the task. So what to do with that? So we will collect data and based on that we can provide recommendations without knowing what will be the students path. Based on historical data we can say the student having trouble here, the student might be having issues in this particular tool or resource. So we can provide that feedback with option for them to cancel the feedback or take it or not take it. So in order to create it we need to collect data from this open-ended learning environment or the TELI. Let us talk about what data you want to collect from this kind of environment. Here is a one example of open-ended learning environment called METAL. It is a modeling based estimation learning environment. It is the learning environment developed to teach engineering estimation to second and third year engineering undergrad students. I will show the video of METAL. This is the screenshot of how METAL works. Please observe carefully what are the tools in METAL, what are the actions student can do, think of that angle. Then let us move on to activity. When the student logged in, we will get the instructions, then you will move on to the problem. The introduction video will show there is a complex problem which can be broken down into smaller sub-problems. So the student can create solution for sub-problems and combine solutions to solve the bigger problem. Here METAL gives the sub-problems for engineering estimation like quantitative modeling, qualitative modeling, calculation, estimation, evaluation. All these sub-problems are needed in order to understand the engineering estimation skill. These each sub-problem is further broken down into a task. You can see that in the right side of the screen. This is called problem map. After solving the problem map, the student has given the real-time problem. So student starts solving the problem. So student solving problem map and selects one particular sub-topic. When you select a sub-topic, you will be provided with a set of questions and answers to answer and also you can ask a guide me to understand how to interact with the system. Or there are other tools like a simulator where student can play the simulator video and try to understand what is going on. And also there is a graph to interact with and these graphs are interactive. Like simulator, we have some other tools say calculator and scribble pad. In simulator, there are two pages. You see the second page now where the learner can change the value of the variables. It is a calculator screen and there is information about different values, parameter values. You have seen the metals screenshot or you have seen the video of metal. Given that, assume that you are a teacher using metal to teach engineering estimation to undergraduate students. What data you will collect about the learners from the metal interaction? Assume that you have skills to collect whatever data you want and your programming skill to store the data, you know what data is to use. Please pass this video and write down your answers. And after writing it down, please assume to continue. You can collect a timestamp of each event and action because time is very important. We talked about it in the MOOC also. Especially you have to collect the learner ID, session ID and you have to store them. You might want to collect the pages, problem, which page they are in, are they in the screen, which screen they are in, which problem map they are in. You can also collect about tools, simulator, calculator, all this information can be collected. What is the student s response to a question asked in the metal and log in and log out information? Also if the students request a guidance or help that can be noted or the student s interaction behavior in the simulator, whether they are increasing the value of the velocity or they increase the value of some other variables, you can store them. In short, you have to collect learners interaction with the system. You can collect all the learners interaction or we can say clickstream data like we discussed in the MOOC, clickstream or trace data. So we can store the data in a database. Here we use MongoDB, no SQL database. Here is the couple of example of data we stored from metal interactions. So in this video, the log data is problem, so the student is in the problem page and from the problem page, the student goes to task and is going to the qualitative model. The object ID is the ID about this particular session and the student ID and the timestamp. So here we have a student ID, timestamp, session ID, then we have a which problem student is currently working on and what is the page he is in. So the page is the action, he is actually doing the qualitative model action, he is doing some hint. The next one, after that the student is still in the qualitative model and he is interacting with the simulator. So now you know that student was in qualitative model, then he looked at the hint, then he is moved to the simulator. In this screen, the student, same timestamp is changing, the student is in simulator, the action is in this page, this log data we can see that the student is still in the simulator, the student moved to the screen 2 in the simulator. Now the action is simulator. So who defines this kind of action, it is us when you create a system, you want to create a action, simulator is action or the page level is action. Now is in screen 2, in screen 2 the student is actually reducing the value of accuracy, sorry acceleration. In screen 2, in screen 2 the student, in screen 2 the student is reducing the value of acceleration. So given this 4 type of data set, we know that the student first went to the qualitative model, then he checked the hint, then he moved to the simulator page. In simulator he moved to the second screen of simulator, in a second screen a student reduced the value of accuracy. So all the students interaction with the system that is clickstream data has been stored in the MongoDB in a time series manner, that is a timestamp tells you that action 1 happens after that second action happens, so in sequence of actions. So in this video we saw how to collect data in a tally or what data to collect. So in short what we are talking is collect all the clickstream data, a student is interacting with the system in a database, SQL or no SQL database. So the format to store the data can be decided, thank you.