Late fame for a project I was involved with at my old uni­ver­sity: The feel­Space belt allows users to “feel” the north direc­tion via tac­tile stim­u­la­tion. It was an exper­i­ment in hack­ing the senses — can we induce new modal­i­ties by technology?

Thanks to Sunny Bains for the nice WIRED story!

Poster here.

Just a short post, but another demo is online.

It is a demon­stra­tion of the “elas­tic list” prin­ci­ple for brows­ing multi-facetted data struc­tures. Click any num­ber of list entries to query the data­base for a com­bi­na­tion of the selected attrib­utes. If you cre­ate an “impos­si­ble” con­fig­u­ra­tion, your selec­tion will be reduced until a match is possible.

The exam­ple data is based on the Noble prize win­ners dataset used in the Fla­menco facet browser.

Elas­tic lists enhance tra­di­tional facet brows­ing approaches by • visu­al­iz­ing rel­a­tive pro­por­tions (weights) of meta­data val­ues by size • visu­al­iz­ing unusu­al­ness of a meta­data weight by bright­ness • and ani­mated fil­ter­ing transitions.

In unfil­tered view, the bright­ness shows a trend mea­sure, indi­cat­ing a ris­ing num­ber of prices of the last years.

In fil­tered views, a brighter back­ground indi­cates a higher weight of the meta­data value com­pared to the over­all distribution.

If, for instance, you click “peace” as in the exam­ple above, you will see that “female” and “Switzer­land” are much brighter, indi­cat­ing that the pro­por­tion of women and Swiss is much higher in this con­text than com­pared to the whole data set. That’s inter­est­ing infor­ma­tion and could also be used to char­ac­ter­ize the result set of a key­word query or any other col­lec­tion in terms of its “char­ac­ter­is­tic” meta­data val­ues. Besides that, it fos­ters under­stand­ing of how meta­data val­ues are cor­re­lated with each other, which is often inter­est­ing infor­ma­tion itself.

You can also switch on lit­tle sparklines to see the tem­po­ral dis­tri­b­u­tion of each meta­data value:

I am cur­rently work­ing on trends in indi­vid­ual tag­ging behav­iour. You might have seen a first, ani­mated ver­sion of my stud­ies based on tag maps. The orig­i­nal ani­ma­tion shows the emer­gence of pre­vi­ously rarely used tags over time. Now I dug deeper and made a richer visu­al­iza­tion for inves­ti­gat­ing this topic.

For the impa­tient: » Check out the inter­ac­tive ver­sion here

And here’s the explanation:

It has been shown before (for a plau­si­bil­ity argu­ment, also check out the mar­vel­lous clouda­li­cious tool, where you can track tag pro­por­tions for any web­site on the web), that tag pro­por­tions for ressources sta­bi­lize over time. Which means that the tag cloud rep­re­sent­ing a tag pro­file for a resource does not change much, once a suf­fi­cient num­ber of tags has been col­lected. In a folk­son­omy, this is gen­er­ally con­sid­ered a good sign, since this indi­cates a cer­tain agree­ment on how to judge a cer­tain ressource and what vocab­u­lary to use.

For tag­ging indi­vid­u­als, and com­mu­ni­ties, this might — at first glance — hold true as well. Con­sider the fol­low­ing the visu­al­iza­tion of a tag­ging com­mu­nity’s evo­lu­tion, for example:

Each tag is assigned a band, with the thick­ness indi­cat­ing the over­all summed usage of a tag over time (time runs left to right). Thus, a ver­ti­cal cut through the graph cor­re­sponds to tak­ing a tag cloud snap­shot at this time point. The ver­ti­cal order is based on the over­all fre­quency of the tags. The color is used to to give an impres­sion of the long tail dis­tor­tion — if all tags would appear equally often, you would see a lin­ear tran­si­tion from red to green instead of the skewed dis­tri­b­u­tion. So — what do we see? Appar­ently, most of the bands seem to grow in par­al­lel, indi­cat­ing a sta­ble growth pro­por­tion for all tags. Of course, we can­not see much for the smaller tags, and there are some edgy parts of the graph which might indi­cate dif­fer­ent behav­ior at spe­cific time points, but over­all — pretty sta­ble impression.

How­ever, this does not make much sense. For indi­vid­u­als and com­mu­ni­ties, the top­ics of inter­est evolve over time, so there must be some hid­den vari­abil­ity not cap­tured by the visu­al­iza­tion and the under­ly­ing lin­ear model.

So I decided to pro­vide an alter­na­tive visu­al­iza­tion for the data based on a decay model, where tags “age” over time and finally get “for­got­ten” if they are not used any­more. This idea is loosely based on the Yules-Simon mem­ory model for tag gen­er­a­tion pre­sented in this paper.

A rad­i­cally dif­fer­ent pic­ture emerges. Not only does the over­all shape now nicely dis­play phases of com­mu­nity activ­ity over time, but also the life cycle of sin­gle tags is much more trans­par­ent. You can rollover sin­gle lay­ers high­light it and dis­play the cor­re­spond­ing tag name. Great fun.

» Check out the inter­ac­tive ver­sion here

What I am now curi­ous about: – Is there a cor­re­la­tion between time-dependency and over-all fre­quency of tags? In other words, are fre­quent tags more evely dis­trib­uted over time, whilst the low fre­quency tags tend to be more vari­able over time? – Is there a cor­re­la­tion between tem­po­ral syn­chro­niza­tion and gen­eral co-occurrence? Which means — do related tags also appear and dis­ap­pear together over time?

I think the answer is YES to both ques­tions, but that would def­i­nitely need some sta­tis­ti­cal analy­sis (any bored neu­ro­sci­en­tists around to help me? ;)

To-dos for the visu­al­iza­tion: – Imple­ment a slider, so you can see how a lin­ear and decayed tag cloud would have looked like at a spe­cific time point. – Sta­men got it right: Maybe I should have plot­ted from the ver­ti­cal cen­ter. Or at least pro­vide an optional inver­sion of the sort­ing. Because right now, all the top (green) lay­ers are really dis­torted, mak­ing visual analy­sis really hard. – Put some num­bers on the axis – Show sin­gle tag­ging events on rollover. Or even “unfold” the layer to improve read­abil­ity and avoid misconceptions.

As promised, here is an update to the tag maps appli­ca­tion I intro­duced below along with some explanations.

For the impa­tient: HERE’S THE LINK

(Update again: The lat­est ver­sion can be found here)

And for the curi­ous: Here’s the expla­na­tions: (more…)

You might have seen the tag clouds posted below. I cal­cu­late tag posi­tions based on co-occurrence, such that tags used together are placed closer to each other. Addi­tion­ally, tags are scaled áccord­ing to fre­quency.
A gen­eral prob­lem I have with the result­ing rep­re­sen­ta­tion (and com­mon tag clouds as well) is the fact, that every tag occur­rence is weighted equally. As a result, these tag clouds never rep­re­sent the cur­rent state of inter­est, but a very slug­gishly chang­ing sum­mary of your archive. How­ever, your inter­ests and the cor­re­spond­ing vocab­u­lary keeps mov­ing on. So I am cur­rently inves­ti­gat­ing trends in tag clouds and how groups of related tags emerge and dis­ap­pear again.

A first glimpse into the dynam­i­cal nature of tag structures.

Just a lit­tle pointer to an ongo­ing project: [edit: » The lat­est ver­sion can be found here «]

I am cur­rently work­ing on sim­i­lar­ity (correlation-based) nav­i­ga­tion mech­a­nisms for tags and other nom­i­nal meta­data and a trend mea­sure (kind of hinted at in the inter­ac­tive ver­sion with the green col­ors). I will soon post an update and a few explanations.

Tag Clouds 5.0 (Ger­man doc­u­men­ta­tion at incom.org)

Inter­ac­tive demo (outdated)

I would be happy if you could con­tribute your del.icio.us, ma.gnolia or furl tags, if you use one of these pub­lic book­mark­ing ser­vices. In this case, just do the fol­low­ing; it’s a one minute thing:

1. Go to https://api.del.icio.us/v1/posts/all?
2. The browser will ask you for your deli­cious user­name and password.
3. As a response, you will get an XML file con­tain­ing all your posts. The browser page might look blank, but if you take a look at the source code, you will see it’s an XML file.
4. Send me the file (copy paste source code into a text file or save directly) along with a short notice, if you want it pub­lished in future exper­i­ments (with your user­name or anonymized).

As a side prod­uct of my work on web feed visu­al­iza­tion, I made a small com­par­i­son of dif­fer­ent ways to deal with tem­po­ral infor­ma­tion in lists of micro­con­tent, such as e.g. blog entries.

(larger ver­sion of the image)

1) Ordered list with­out gaps: Clearly, the most space-efficient solu­tion — how­ever, only tem­po­ral order­ing is pre­served and not tem­po­ral struc­ture. It is not visu­ally evi­dent how the items are dis­trib­uted over time.

2) Cal­en­dar: Each time unit (days for exam­ple) has equal space assigned, regard­less if there are items assigned or not. A pre­cise dis­play, how­ever, very space-inefficient, since a lot of the dis­play space is typ­i­cally used for dis­play­ing “nothing”.

3) Accor­dion: Sim­i­lar to cal­en­dar view, but empty time units are dis­played on much less screen estate. This gives a pretty good first-glance impres­sion of large gaps and close-together items. How­ever, depend­ing on the tem­po­ral struc­ture, there might still be large streaks of wasted space for large gaps.

4) Folded gaps: This is the solu­tion I pro­pose (and which I believe is novel. If oth­er­wise, I would be happy about a short notice!): Tem­po­ral gaps are dis­played as if a part of the list was folded to the back of the dis­play. Short gaps have almost the same size as in accor­dion view. Long gaps are larger, but do not grow lin­early, but with the square root of the num­ber of empty time units. Visu­ally, this is jus­ti­fied by intro­duc­ing shad­ing to indi­cate that the “orig­i­nal mate­r­ial” is folded to the back. Fold­ing also pro­vides a plau­si­ble model for inter­ac­tive adjust­ments such as reg­u­lat­ing the gap size.

To sup­port my argu­ment, I also made small demon­stra­tor based on actual web feed data. It takes a while to load (~700k of data), so please be patient. On the left, you have a menu for select­ing dif­fer­ent feeds. On the right, I drew a con­nec­tion of each item to a cal­en­dar with fancy curved lines. You can adjust the size of the dis­played items with the zoom slider.

Let me know if it works for you — tech­ni­cally and conceptually!