Once upon a time in the far far land of… as a first time blogger it’s hard to think of an interesting and eye catching beginning and as all fairy tales start in an undefined time and space so does this post.

Not long ago (the time), I participated in a PROGRESS course (courtesy of Nano2Life) in Chichiliane, France (and space – which the castle we stayed in reminded me of a fairy tale) and one of the first lessons we were thought was how to give a good and practical feedback to a person (I-message). The I-message should include (among other parts) the person’s behaviour and NOT the person’s giving the feedback interpretation to that behaviour (e.g. “You didn’t do your homework” and not “You are lazy”). The point to this feedback is to be very specific and not to try to give more complex explanations/perspectives to the human behaviour (that is not the goal of a feedback). Can we really do that? maybe when we want to give a good feedback we should stick to the ”recipe”. Human behaviour is so complex and represents our most deep (and also shallow) thoughts and emotions. As you can “see” I’m very interested in human behaviour but more to the reasons people act the way they act and more specifically to the cognitive aspect of human behaviour from the educational point of view (how they think).

In 1956 Benjamin Bloom developed a system of categories of learning behaviour to assist in the design and assessment of the educational outcomes in three domains: Cognitive, Affective and Psychomotor (best known as “Bloom’s Taxonomy Of Educational Objectives”). Although Bloom described a taxonomy for educational objectives and it was used to plan curriculum and lesson plans, in the cognitive domain, Bloom tried to bridge between the student’s learning behaviour and his/her level of thinking (e.g. the level of “synthesis” can be described by the following student’s learning behavior: develop, plan, build, create, design, organise, revise, formulate, propose, establish, assemble, integrate, re-arrange, modify, etc.) and emphasized the thinking skills and processes and how they can be manifested in the learning behaviour (in learning environments).

The use of web based learning and teaching environments (WBLE) is growing in a rapid pace (as does research on the use and effectiveness of such environments). As researches, teachers and developers we need to get to know our students and more interesting for me, how they think and what processes they undergo when they are learning online. Web based learning assessment tend to focus on the outcomes and not the processes, especially not the thinking ones. And here is where the use of Educational Web Mining (or Educational Data Mining) enters the scene.

Using Web Mining to trace the students’ behaviours during their learning process can help us (researchers, teachers and developers) to better understand their thinking skills and procedures they are using in solving different problems, or in other words, how our mind is manifested in the logs (WebMinding). How? well, this is not as easy as it sounds but Bloom’s seminal research is a good beginning. Can we interpret behaviour to thinking without seeing the students? Can we translate the behaviour to more complex aspects of the learning process? That is what we are trying to do. This Blog (as for the others I think) will help me clarify different issues “out-loud” (using meta-cognitive skills) with myself and discuss those issues with the big outside world.

For years, I am developing curricula materials in science education for elementary school students. All the materials are on a Web site, in which students can learn, play and take exams. In many aspects, I am their second science teacher, although I never saw them, and they never saw me…  It is always the same scenario: after a team meeting in which the educational objectives of a specific activity are discussed, there is me and the content left. That is the time when I am beginning to struggle with my self with questions like: what is the most important information that should be presented? How to present it? What kind of activity would better suit teaching this knowledge?

And then I am thinking about the kid behind the computer, the kid that I’ll never get the chance to see… and I wonder: Will he like the educational game that I have developed for him? Will he read the informative text within the activity? Will he be able to understand it? or maybe his low motivation will lead him to skip to other districts on the Web whenever he gets the chance?

A lot of questions are rising, and all of them are draining to two main questions that are interested every instructor who engaged in online education: What the students learn? How they learn it? It is clear to me that the first step for answering these questions is to know my “customers” better then I do now.This situation, in which the developer of the educational materials does not meet his students, is similar to the situation in which a tailor needs to sew outfits for customers that he never saw. Will he succeed to sew them outfits that suit their exact measurements?

From time to time I am visiting classrooms to observe students learning with the materials that I have developed. But, however, this way I will never get the chance to see the whole picture. I will never reach all the students – all my “customers”. And more important – I will never understand the learning process they are going through.

And now we have the educational data mining.

By data mining, every single action of a user can be documented and stored in computerized process. The footsteps that the students are leaving on the educational site, make it possible to observe learning processes and strategies, recognize difficulties and barriers, examine preferences of students, analyze behaviors of students and so on.

The use of data mining on educational sites to understand how students learn is, of course, more complicated than using it on commercial sites for trying to understand the characteristics of customers, because thinking processes are involved. Yet, it opens great opportunities for the online developer to learn about his students and their interaction with the contents and tools.

He (the developer) can get information such as: What kind of activities are causing great motivation? Who are the students that have great motivation? Why them? What kind of scaffolding should the developer supply to students? What kinds of learning behaviors leading to success in specific online activities? How to suit an activity to a specific student? And so on. Therefore, the data mining enables the students to become “closer” to the developer, and influence the development process by helping to suit them the exact “outfit” that they need.

For almost four centuries, Descartes’ “Cogito, ergo sum” lays as one of the humanity’s foundation stones (and it is clearly a non-understandable one). Therefore, it is of no surprise that when today’s new thinkers are asked about their motivation, they simply answer: “Blogito, ergo sum”. It is quite awkward that stamp collectors, rock climbers, Second Life gamers, persinds of huge countries and others who chose a strange hobby are not struggled with the “Why do you do it?” question as often as bloggers.

I must admit that I havn’t yet found my answer to that question, even after a year-and-a-half of blogging about my life as a young father, and after a year of being a double blogger (the second blog is about my genealogy hobby and is also a joint one). So, here I am now, a triple blogger with no answer to the million-dollar question. But, from the other hand, if there was a simple answer to the million-dollar question, it wouldn’t worth million dollars, would it?

However, blogging about my research area seems quite natural to me. Academia-related discussions within world-wide research communities take place mainly in journals and in conferences. It seems that the Internet dramatically changed the way scholars read journal articles and organize conferences, but did it really change the way we exchange ideas outside of our everyday immediate neighborhood?

It might be true that most of the discussion regarding one’s own research occurs between him and himself, while he’s reading, writing, thinking, eating, taking shower or even doing shopping. It is almost a continuous self-discussion. It will take months until this person wil be able to share his thoughts publicly with world-wide colleauges in the next conference, and if those thoughts will eventually be formed as an article, it will take extra more months for the rest of the world to read it in the journal (in case, of course, that the rest of the world have access to that journal and that they actually know about or accidentally find this article). But sometimes, it’s really shame not being able to share the thoughts themselves, since they may consist of so much fruitful ideas and so much future paths to be followed. These paths might be very interesting to discover if only were talked.

In our research area, it seems even more natural to document the thoughts: we deal with other’s traces, but what about ours? We aim on analyzing students’ hidden information from continuously collected raw files, but our hidden not-spoken and not-bitted continuously gathered thoughts are often not documented at all.

This blog (at least for me) will be the place to share thoughts regarding anything related to Web (data) mining in education. I hope it will help me to understand myself a little bit more…

Twenty-five years ago, while I was a Ph.D. student in Tel-Aviv University, we developed our first educational software on the Apple II personal computer. This software, PolyLine, allows students to explore variables that determine a pattern drawn on the screen, learning the skill of controlling variables in scientific experimentation (you can try it in a modern Java demonstration). Being curious to learn about the user’s learning process we embedded in the software a tracing module that recorded each one of the user keystrokes. It was my first time to analyze automatically-collected data, and I was fascinated by the quality of conclusions that can be drawn from it and from the ease of collecting this data. This was the first time I edumined.

Later on, when I started my career as an educational researcher my students and I frequently used data mining in our research on computer-based learning and I became more proficient eduminer. But it was only during the last decade, since the Internet usage became as popular as drinking water, that I really realized the potential of Web mining for education. I started to understand how much we can learn about the learning process by looking at the traces learners are leaving in the servers of the e-learning application in a form of Web logs. All is needed is an educational vision and a set of tools and techniques that will extract the Web logs data and translate it to learners’ behavior variables. This is what our research group in Tel-Aviv University is doing now. In this newly-born blog we would like to share with you our inspirations and dilemmas, insights and frustrations, as well as some inside stories beyond our work. We welcome you to our blog and hope you’ll become an active participant in our edumining community.