Back from Easter holidays I discussed with Patrick this cumulative map representation. The goal is to make visible the part of the map that attracted more of the user’s attention or time. For this reason I was trying to count the number of items that fall into a sub-area of the map; how many of these items were read by the users and how many selected by the users.

One of the possible visualization strategy that we discussed is the idea of using ‘heatmaps‘ a kind of bi-dimensional gradient that can highlight hot spots on the map (they are particularly used in cognitive science to represent eye tracking).

Another idea that we discussed briefly is that of showing the sequence of action on the map. In this way will be eventually possible to detect if there are common pattern of usage among the users (i.e., how many moved from square 1 to square 2?, and so on).

For the moment I am stuck with a minor visualization bug …

Tags: clustering, information visualization, map algorithms, maps, spatial clustering, statistics

I have been pocking around to generate a visualization that can help understand how the users interacted with the map in our virtual retrieval experiment. It was fun to discover that there were some zones more used than others in the map, namely the central upper quadrant and the central left and right quadrants. The image below shows the density of visualization of results.

The map has been subdivided into 9 quadrants. In each quadrant is possible to see three big numbers. The upper one is the number of results that fall into the sector (even if you see few points there might be super impositions between similar queries). The lower-left big number in blue is the number of the items that have been read by the users. The red is the number of items that have been selected by the users in that quadrant.

Tags: information metric, information retrieval, information visualization, Latent Semantic Analysis, map algorithms, maps, search engine, spatial clustering, spreading activation

Dear Reader,

for the development of my thesis’ work I prepared, together with Lorenzo Viscanti, an experiment on Visual Information Retrieval. I ask your participation in this experiment that will last not more that 15/20 minutes.

If you agree, you will need to visit the following web page: http://www.noosfactory.com/visual%5Fir/ . There you will find some more information on the assignment: you will be presented two tasks, each of which will last 5 minutes. Each task will require you to run some queries on a collection of news articles compiled by Reuter during the ’80s. Finally, you will be posed 4 brief questions.

Please note that there will be an extra 5 minutes to download the application that you will use to participate in the task. This time may vary depending on your internet connection speed.

Tags: information retrieval, Latent Semantic Analysis, map algorithms, spatial clustering, spreading activation

Recently I managed to calculate the Minimum Spanning Tree (MST) of the Urban Tapestries dataset that was collected during the trial of 2004. From Wikipedia:

Given a connected, undirected graph, a spanning tree of that graph is a subgraph which is a tree and connects all the vertices together. A single graph can have many different spanning trees. We can also assign a weight to each edge, which is a number representing how unfavorable it is, and use this to assign a weight to a spanning tree by computing the sum of the weights of the edges in that spanning tree. A minimum spanning tree or minimum weight spanning tree is then a spanning tree with weight less than or equal to the weight of every other spanning tree. More generally, any undirected graph has a minimum spanning forest.

Copyright notice: the present content was taken from the following URL, the copyrights are reserved by the respective author/s.

For the Urban Tapestries dataset, I had to adapt a version of Kruskal’s algorithm for minimum spanning trees that was greatly implemented by Prof. David Eppstein (thanks!). So the final result is visible below. Now the next step is to calculate a MST over the semantic distribution of the points and then find a way to compare these two trees to measure the distortion of the two layers.

Copyright notice: Google 2005, The GeoInformation Group 2006, the copyrights are reserved by the respective author/s.

Tags: anamorphosis maps, beta-skeletons, clustering, information metric, information visualization, map algorithms, maps, spatial clustering

R. M. Ripley, A. Sarma, and A. van der Hoek. Using visualizations to analyze workspace activity and discern software project evolution. ISR Technical Report UCI-ISR-06-1, University of California, Irvine, California, USA, 2006. [pdf]

———————–

In this paper, the authors presented a prototype of a 3D visualization for workspace activity. This visualization shows only active workspaces and artifacts and depicts activities as changes to artifacts. Each change to an artifact is denoted as a cylinder and stacks of such cylinders represent activities in a particular workspace (developer-centric) or activities carried on specific artifact (artifact-centric). They applied the visualization to several open-source projects and demonstrated pro ject evolution and other interesting situations.

The visualization has two primary modes: developer-centric and artifact-centric, which we will discuss shortly. Common to both modes, the visualization shows only active artifacts and workspaces, thereby eliminating the clutter of inactive entities. Stacks of cylinders represent workspace activities

(changes to artifacts), with each cylinder corresponding to a particular artifact in a particular workspace with the dimensions representing the size of the change (the bigger the change, the larger the cylinder).

In the developer-centric view (see the figure, inset), a stack of cylinders represents a developer’s workspace with each cylinder representing an artifact being changed in that workspace. Workspaces with many activities correspond to tall stacks of cylinders. The stacks of cylinders with the most recent changes are placed in the front of the view and, as time elapses, stacks for workspaces with “older” activity slowly start moving to the back, representing dormancy. Thereby a user is able to quickly discern the loci of activities from the height of the stacks and recency of these activities from their position. The artifact-centric view (see the figure, bottom) behaves similarly to the developer view, but instead each stack of cylinders represents a particular artifact and each cylinder in the stack represents changes to that artifact made by a particular workspace.

Copyright notice: the present content was taken from the following URL, the copyrights are reserved by the respective author/s.

Tags: collaboration tools, information visualization, software visualization

Trying to help Nicolas to find nice symbols for the CatchBob! semi-structured interface, I  recalled this nice web site that I visited already in the past. It’s a visual dictionary of symbols. For each pictogram it shows references that uses the symbol and a description of the cross-cultural meaning. Different search tool are available.

Tags: information visualization, universal cognitive distance

The last 16th of October (2005) a participatory simulation was organised a Toulouse during the Science Fair for the World Year of Physics. The experience simulated the famous Brownian movement, that was theorised by Einstein hundred years ago.

In this experience, the particles with random movements have been represented by high-school students. They moved randomly in the city center choosing randomly the direction of their movements. In the following map is possible to see a representation of the particles’ position after 50 crossing.

Results showed a distortion of the obtained results from the expected results (i.e., the position of the center of gravity is shifted on the right hand-side on the map). A further exploration of the data allowed to explain this shift by the presence of an architectural barrier on the city path. The river, in fact, played an important role in directing the direction of the particles on the right upper quadrant of the map.

This is in relation with my previous post where I was highlighting that geography and geometry are not synonymous. We live and we perceive a geographical world, which we try to model with geometrical models. These models are often limited as not incorporating important factors as, in this case, natural barriers.

Tags: field tools, map algorithms, maps, mobile learning, psychogeography, participatory simulations, urban exploration

Podcasts are rapidly emerging as a popular on-line audio publishing vehicle.  But trying to find podcasts that are of interest to us is reminiscent of the early days of the internet.  Instead of using general search techniques, sites attempted to organize all the content on the internet into categories.  The diversity of the available content became too cumbersome for mere categorization to be an effective means for people to find what they wanted.  Search evolved to where it is today, that is, to look inside websites quickly for words and phrases.

Podcasts have been subjected to the same primitive search through categorization … until now.  PODZINGER looks inside podcasts, not just the metadata, letting you search podcasts in the same way that you search for anything else on the web.

When you type in a word or terms, PODZINGER not only finds the relevant podcasts, but also highlights the segment of the audio in which they occurred. By clicking anywhere on the results, the audio will begin to play just where you clicked.  There are also controls that let you back up, pause, or forward through the podcast.  Or you can download the entire podcast.

Tags: search engine

Deciding how much latitude to give users is an essential part of any mass-customization program, says B. Joseph Pine, the author of Mass Customization: The New Frontier in Business Competition (Harvard Business School Press, 1993). “Fundamentally,” he says, “customers don’t want choice. They just want exactly what they want. Your job is to help them figure out what it is they want, because often they don’t know or can’t articulate it.”

That’s in essence at the core of LEGO new business model that articulates in 3 steps: (1) the user chose the design s/he wants using a support tool; (2) The factory creates an ad-hoc packaging of the bricks; (3) the packaging are sent to the customer.

I am not sure this model fulfill the idea of helping the user to figure out what s/he wants but I really like the idea of customization of the bricks set.

(via)