Visualizations in the feedback loop

Second post of the day: I surprise myself!

This evening I attended to the workshop of R given by @jfelipe (Felipe Ortega) corresponding to the Visualizar'11 seminar. I was playing a bit with R in the past, and for me is not much more than the "open Matlab"; but it has nice features like extensibility installing new libraries and the graphs obtained are pretty beautiful... but interactiveless!

But after the workshop I had the great opportunity to talk with @infosthetics (Andrew Vande Moere) about my own research, that is, my PhD, and how I am introducing now the concept of visualization in it. And the 15 minutes we discussed were much inspiring and made me reflect about the directions that my PhD is taking (yeah, it's kind of live creature). Let's try to summarize the main ideas.

One of my previous ideas we discussed was the visualizations of learning events collected in the lab sessions will help teachers to make better decisions: visualizations to make them more intelligent (as stated in my previous post). Teachers will be able to orchestrate the classes in a better way and provide more effective and quick feedback to the students (hopefully!). I confirm that this is a good idea if executed appropriately: cool!

But the majority of the discussion was about the other leg of my topic: introducing visualizations for the students for awareness and reflection. He recommended me three important things these visualizations should comply with to be engaging (I have to measure it!!) for students: they have to be beautiful (esthetically good), fluent (that is, very straightforward and simple interaction, like this example) and that is part of a story. As I understood (if I did at all), this means that the visualization is not the end of the process (of learning) but a part of the story; it has to put into the mind of the student the next action he/she has to take (yeah, this is getting complicated!). Anyway, I am going to present to the student these engaging visualizations with their individual data and the aggregation of the data of all the students in the class. And I am going to collect all the data I could about access to the visuals, discussions between students and with the teacher, etc. to be able to demonstrate engagement.

And then I have to evaluate the tool I am developing, in a real scenario: typical control group and experiment group. I have to think better how to measure engagement (qualitative parameter measured as quantitative as possible?) and learning, decomposed in the attainment of learning outcomes that the learners are achieving during the course. How to measure that? Think, think, think. And, of course, I have to have always in mind the ethical implications of my experiments: one of the groups will have an advantage over the other if testing my tool with them? What if I don't know exactly if it works? I feel the responsibility over my shoulders...

Thanks Andrew for the nice+inspiring+engaging+fascinating conversation! I very very much appreciate it!

Visualization as a new keyword

During the last week I have been learning a lot about data visualization, above all thanks to visualizar11. Let's reflect a bit about the experience up to now...

Visualizar'11 (subtitled "understanding the infrastructures") is a workshop lasting 3 weeks (from 14/6 to 30/6) intended to develop several projects related somehow to data visualization with the underlying leitmotiv of the infrastructures. The first two days were devoted to keynotes and "communications" from experts in the field, and since then you can participate in one of the 10 selected projects (in my case, in order to learn a bit more about the topic). Right now I am more or less implicated in a project related to visualize the Google technological infrastructures in order to show implications to the "don't be evil" stuff: open source and closed components, effort devoted to them, etc. Let's see where it takes...

But why (the hell) do I want to learn more about data visualization? Because visualization is one of the tools for analysis of collected learning statistics (events). If well used, it can be very powerful since it permits to find relations among data not directly related to other data. As Enrico Bertini said in Visualizing Europe last week, we have to find the concrete field or context where visualization was not useful but indispensable. And I think that for learning analytics in class (in the enactment phase) it is indispensable for facilitating teacher monitoring of the activity of the students (in the context of a lab session).

W. Ross Ashby, from wikipedia

Another concept I want to highlight is Intelligence Amplification (W. Ross Ashby, 1956) as opposed to Artificial Intelligence. As Inês Salpico so well introduced during her communication at Visualizar11, the visualization is not supposed to be an AI product but a IA tool, that is, a tool that helps the user to amplify his/her potential (let's say intelligence) in order to accomplish a concrete task. And this is actually what I want to do in my PhD!!

Will I enable the visualization of the learning infrastructures? Hopefully!