Tuesday 13 November 2007

Temporal Information Visualisation

Hans Rosling, a professor in Public Health from Sweden, shows how you can take temporal information to make sense of your data - in the cases below the relationship between developed and non-developed countries. He comes up with answers which might surprise you. This is his 2006 TED talk:






Now for his 2007 TED talk (watch out for his sword trick at the end - had me cringing!)





I blogged about the relationship between information visualisation and information retrieval. What amazes me about these presentations is that Hans is able to present very complex data in a way which is very easy to comprehend for a non-expert. Case of a 'visualisation intermediary' at work here?

Mobile identity

There are a number of things I like about this talk by Jan Chipchase including the concept of mobile identity:





It made me wonder - how could we use these ideas for mobile search? (no I don't have any answers).



UPDATE:

Jan Chipchase talking about design for illerate users:



Sunday 11 November 2007

Creativity Stifled

Larry Lessig talks about how creativity is being stifled by the Law. Its a very wide ranging talk and I can strongly recommend it.

I've been posting about DRM recently e.g. here and here and while Lessig's presentation tackled copyright in general, it is clear to me that the thrust of the argument is against technologies such as DRM. It criminalises people for no good reason.

The argument used is lets go back to what you used to do before the advent of technology such as the phonograph (record player to oldies like myself, or CD's to you younger folk). Musicians earned their living by traveling around and giving performances. Music was a shared experience. Lets go back to that.

The danger is not restricted to just this one area - stifling of creativity in the software area is another bone of contention with me. Software patents are stupid, and should be abolished. They are virtually only ever used (apart from a tiny minority of case) by companies to counter sue other companies, who might sue them for a copyright breech. So the only reason to have patents in software is to make lawyers rich. erm...

Geras vs. Kamm on Blogging

Further to my post on New stories and Blogs, Norman Geras (who writes in defence of blogs) has kindly posted an article on Aquinas on blogging which gives a link to various posts in which he defends blogs (political blogging in particular):



I'm with Norm on this one, his arguments are much more persuasive. Kamm uses a political blog to put forward his arguments, so one can assume that the arguments he uses against political blogging do not apply to him - very much a case of special pleading. If my assumption is false, why does Kamm write on a political blog? [Note: Kamm is not against all blogs, he makes this clear in an update at the end of this article.]



This isn't to say that Kamm's views are totally off beam - I think there are instances where (as Geras points out) the behaviour of some bloggers and commentators to blog is nothing short of appalling (e.g. an anti-Semitic post on Cif which I will not link to even if I had time to find it again). Norm points out that this is a failure not so much of blogging, but the standard of behaviour expected on them (a lot less than say a public meeting).



But isn't this an argument of free speech? If an argument on a blog is poor, relying on poor or no evidence, can we not use a blog to shoot it down? I quote Geras "And no one is forced to be a consumer" - you are free to ignore blogs.



More links to this issue:





Footnote: the post was in response to is Philosophy good for the soul?, which is where the reference to Thomas Aquinas in Norms blog comes from.

Reflections on Tone and Voice: paper by David Huron

In a music informatics reading group I attend when I can, we discuss a paper by David Huron, entitled Tone and Voice: A Derivation of the Rules of Voice-leading from Perceptual Principles. Here are a few random thoughts and reflections arising from my reading of this paper.



Much of it unfortunately went right over my head, largely due to my ignorance of some important music concepts. However...



The toneness principle in my view may well give us an idea on music relevance e.g. how users may perceive and identify music. This is image identification with sounds or as Huron puts it 'sounds may be regarded as evoking perceptual images'. Initially I interpreted this at rather a high level, but the reading group pointed out to me that actually the sounds are a low level e.g. a passing car, bells ringing - in other words single sound. Which doesn't in any way go against my argument. It would be an interesting project to see how or if users do in fact associated sound with images (if it hasn't already been done) e.g. the way I associate Beethoven's pastoral symphony with the countryside or Dvorak's new world symphony with the North of England (Hovis ad). In my view high or low level identification of sound with images could provide us a link to music relevance.



Later on he talks about the principle of temporal continuity, which further reinforced my views. Huron states 'auditory images have a tendency to linger beyond the physical cessation of the stimulus'. The examples he gives examples of timpani rolls or bubbling brook. A sound invoking an image which stays with you. More evidence of image identification with sound if not music? Perhaps the length of sound of the instrument may have some influence here (the longer the sound, the more likely it is to invoke an image to the user).



Perhaps music can evoke motion as well as images. Huron states that 'it would seem that the sense of continuation between two tones is an auditory analog to apparent motion in vision'. The example he gives is some research using two lamps that could be switched on and off. The 'sense of apparent motion depends on the distance separating the to lamps and their speed of switching'. If may me speculate on the effect two instruments (say) moving in and out of voices could invoke movement, and hence contribute to our understanding of music relevance.



In Fig. 17 of the paper a classification of music is given based on a two dimensional graph of onset synchronisation vs semblant motion (melodies moving in the same direction). It was speculated in the reading group that you could use this as a filter for query by example e.g. pick collections of music in some space on this graph and then use user generated knowledge (via meta-data) to further refine the query.



And now for something really quite enjoyable. The Huron paper mentions the Hurdy Gurdy, an instrument I'm quite interested in, and I found this video of a performance by Melissa Kacalanos on the New York subway:





Marvellous, truly marvellous! She has her own Web page. This however is just bloody funny! What this man can do with his hands is nothing short of amazing.





LOL! Bohemian Rhapsody will never sound the same again. He actually gives a tutorial as well:





Have at try....



References


Huron, D. (2001). Tone and Voice: A Derivation of the Rules of Voice-leading from Perceptual Principles. Music Perception, 19(1) pp1-64.

Friday 9 November 2007

The death of DRM

Further to my post on Music IR, it appears that DRM is not likely to be around for very much longer. Some evidence:




  • DRM free music sales outnumber DRM sales by 4:1.

  • DRM free music sales shifts albums - 70% MP3 sales being full albums.

  • Shift from unbundling albums to single sales caused revenue decline



I will not be mourning the death of a pointless and self defeating technology.

Monday 5 November 2007

The Day The Routers Died

I wonder how Don McLean would feel about this one?





A LOL moment...

Sunday 4 November 2007

Reflections on the relationship between information visualisation and retrieval

I’ve been thinking about this for a while now, but not really got a handle until recently on how to progress this area of research forward. When examining a PhD thesis not long ago, some thoughts on Geo-visualisation specifically and Information Visualisation in general, and the relationship with ASKS (Anomalous States of Knowledge) were brought to the fore, particularly with respect the underlying task. My thoughts were brought into sharper focus when attending a Geo-visualisation tutorial run though more recently.

Why? Lets think about ASKS in terms of information retrieval. A user has a certain level of knowledge when approaching search, which will have an impact on how well they will proceed with the search process and how long it will take them. Lets say we ask two people to search for the effect on hydrogen atoms of the heat generated internally by the Sun. One person is a scientist the other is just someone off the street with a weak knowledge of science. Its clear who will do better in the search process here (no prizes for guessing who).

The task itself has an impact on the search process. A Professor who searches for articles on ‘the effect on hydrogen atoms of the heat generated internally by the Sun’ for background knowledge to their own research will probably be looking for a very different set of documents to a student who wants to write an essay on the same subject, set as coursework by the professor. The student will want more articles with fundamental principles in it than the professor, who is more likely to be interested in cutting edge research.

The information source and the users understanding of it play a very important role on the search process. Our professor above will have a much better handle on the sources than their student.

Ability to understand and use the software is yet another component. How complex is the interface? How well does the user understanding Boolean logic operators to carry out sophisticated searching? A case were an information scientists with knowledge of understanding users information needs and expertise in turning these needs into a search will have a distinct advantage here, particularly if they have been using the software for a while. The students I teach often do these roles, as ‘search intermediaries’.

OK, were are we going with this? What’s the link to information visualisation/Geo- visualisation?

First lets deal with the knowledge which a user brings to the visualisation process. It’s clear that the process will be undertaken with a certain amount of prior knowledge of the domain, whether its visualising maps or multi-dimensional data.

The task assigned has an impact. A professional geographer will be doing very different things to a student who is learning about the process of Geo-visualisation. The profession geographer will be perhaps trying to learn new things in the visualisation process. The student however, is more likely to be learning about things which are well known, and which an educator (perhaps the professional geographer) can help them with. This really came out in the PhD I examined.

The source data as an impact. Our profession geographer may well be working with more complex underlying data than our student trying to understand Geo-visualisation. You’ll be using different data for the different tasks I think. This really came out in the PhD I examined.

The software used has an impact. A Geo-visualisation expert with expertise in the field will have abilities in using the software, which could put them in the kind of role described above in IR – as some kind of ‘visualisation intermediary’. This stuck me in the Geo-visualisation tutorial I attended.

I think there are similarities between information visualisation and information retrieval that I think are worth investigation, in particularly the many information seeking models which are around e.g. Wilson, Ellis etc. Seems to me that IF is a seeking process, and we might be able understand it a little better if we apply these models. I use ASKS above, but information retrieval is not just about problem solving.