Dear All

I have just been over the past two days to a Information Visualisation conference over here in London. Along the many sessions and presentations about various topics of visualisation, the GeoAnalytics Session I found the most interesting.

A particularly impressive presentation was given by a research group from Sweden, Linkoeping University. The VITA research group there focus on developing advanced tools for the visualisation of large and high-dimensional (meaning 100+attributes) spatial and non spatial datasets.From their research, they have developed a toolkit called "GeoAnalytics Visualization (GAV)", which allows users to dynamically explore time-varying, geographically referenced and multivariate attributes. These GAV components are developed in C# based on Microsoft’sDirectX graphics library and are freely available for download and use on their website.

Some of the main characteristics of the tool are listed below:

• Layered component approach for customization, scalability and reusability;
• Shorten development time by utilising already developed and assessed components;
• Mechanism for integrating external user components;
• A 3D data model for spatial-temporal and multivariate attribute data exploration;
• Hardware rendering techniques using DirectX for interactive performance;
• Visual space-time and multivariate inquiry tools;
• Component-embedded colouring, interaction and filtering;
• Integrated mechanism for saving and packaging the results of a VA reasoning process;

(for more information and demo applications, you can go to the groups website http://vita.itn.liu.se/gav . (Altough I couldnt get their demo applications to run, my Windows XP throws error messages. if anyone manages to run the application, would be glad to know whats wrong with my PC).

I was really struck with the great visual capabilities for data exploration, analysis, clustering and querying this toolkit provides. It is hard to put into words the capabilities of the system, a demo video such as you can watch on this site will explain this all much better. ( see http://vitagate.itn.liu.se/projects/GAV/demovideos/zipViewNewProject/zipViewNewProject.html). Also of note is the speed with which zooming, panning and thematic formatting was possible in the demo application, I dont know about the circumstances, and how this can be compared to Manifold's performance, but it seems that they make use of GPU hardware graphics acceleration to speed up the visual interaction.

How does that apply to Manifold?

Given that Manifold at present lacks much advanced Exploratory Data Analysis tools, specifically visualisation tools that present multiple linked alternate views of data, how could we Manifold users leverage such a toolset?

Given that this toolkit is freely usable and is developed using Microsoft .Net, how easy would it be to create some tight coupling of sorts to integrate alternate linked data visualisations using this toolkit inside Manifold?

What are the prospects for Manifold itself to develop capabilities in exploratory data analysis and visualisation? Are users here even aware of Information Visualisation and its capabilities, and do they need such capabilities? There has been previous discussion of requirements in the following thread, but feedback from Manifold has been sparse on the topic. (see http://forum.manifold.net/forum/t55699.31)

I hope some of you might find this of interest and worthy of discussion.

Cheers

Patrick

Check out the Manifold Wiki with SQL and scripting examples at http://www.manipedia.eu/

Spatial Knowledge, my personal blog.

"Given that Manifold at present lacks much advanced Exploratory Data Analysis tools, specifically visualisation tools that present multiple linked alternate views of data, how could we Manifold users leverage such a toolset? Given that this toolkit is freely usable and is developed using Microsoft .Net, how easy would it be to create some tight coupling of sorts to integrate alternate linked data visualisations using this toolkit inside Manifold? What are the prospects for Manifold itself to develop capabilities in exploratory data analysis and visualisation? Are users here even aware of Information Visualisation and its capabilities, and do they need such capabilities? There has been previous discussion of requirements in the following thread, but feedback from Manifold has been sparse on the topic. (see http://forum.manifold.net/forum/t55699.31)"

I hope some of you might find this of interest and worthy of discussion."

Patrick,

Very interesting. I would not have minded going to that meeting myself. Impressive work.

I definitely agree that Manifold should have such capabilities and probably then some (it is not too easy to tell from the website what the full range of GAV's capabilities is). However, no matter how tempting, i do not think it would be a good idea for Manifold to "tightly integrate" with GAV, unless tight integration means invisible and seamless from a usage perspective. I also think it would also be unwise as it would make Manifold a hostage to the development of a third party product, even if it is as powerful as it looks. The best solution IMO would be for Manifold to develop its own similar features ito those in GAV and then some, using GAV and others as a model.

Detailed discussion about scientific information visualisation and analysis would be a good subject for intensive and detailed discussion in London next February.

Dream no small dreams for they have no power to change the minds of men. - After Johann Wolfgang von Goethe

Nick

I am not advocating that Manifold should depend on external toolkits to provide "core functionality".

I was more interested, from the standpoint of a .Net neophyte, to know from some more seasoned .Net developers the feasibility of integrating this toolkit with Manifold. Be it through a toolbar which launches different linked visualisations linked back to the Manifold dataset, or an external application which can read Manifold project files? I dont know what is possible, or makes sense in this context.

Obviously, if Manifold want to be serious about developing such functionalities, that would be appreciated as well? GAV certainly is an impressive model to get inspiration from.

Check out the Manifold Wiki with SQL and scripting examples at http://www.manipedia.eu/

Spatial Knowledge, my personal blog.

"...the feasibility of integrating this toolkit with Manifold."

This is what i fear. I fear it because it , could be used as an excuse for Manifold not developing these kinds of native capabilities, which would be a shame and a bad move IMV. However, I am optimistic that would develop them given enough encouragement and momentum.

I would like to know more about GAV can do. Because, all there is to go on ATM is a few brief decriptions and some graphics. From what i can judge most of what is described is pretty much mainstream in the HPC scientific visualisation community, albeit mainly done on supes. Doing web searches using the terms scientific visualization (with z), high performance visualization, high performance computing will come up with loads on this subject. A lot of the algorithms and concepts were developed by physicists, engineers and medics.

Dream no small dreams for they have no power to change the minds of men. - After Johann Wolfgang von Goethe

Patrick,

Saw your post this morning - downloaded all the examples and have been playing all day - cracking stuff!!!

Translated everything to VB code and now have to figure a way of accessing Manifold - it's exactly what I need at the moment

Well spotted!!

Jonathan

Hi Patrick,

Very interesting tool. I have been researching and experimenting with some of these tools for many years. Another very interesting tool to explore is the GeoViz Tool kit (http://www.geovista.psu.edu/geoviztoolkit/index.html).