From angus.forbes at gmail.com Thu Jan 12 17:09:34 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Thu Jan 12 17:09:40 2006 Subject: [Visinfo] Visinfo mailing list test Message-ID: testing... From nsbe_pr2000 at hotmail.com Thu Jan 12 18:37:51 2006 From: nsbe_pr2000 at hotmail.com (Reginald S. Archer) Date: Thu Jan 12 18:38:01 2006 Subject: [Visinfo] Visinfo mailing list test In-Reply-To: Message-ID: works -----Original Message----- From: visinfo-bounces@zydeco.mat.ucsb.edu [mailto:visinfo-bounces@zydeco.mat.ucsb.edu] On Behalf Of Angus Forbes Sent: Thursday, January 12, 2006 5:10 PM To: visinfo@zydeco.mat.ucsb.edu Subject: [Visinfo] Visinfo mailing list test testing... _______________________________________________ visinfo mailing list visinfo@zydeco.mat.ucsb.edu http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo From angus.forbes at gmail.com Fri Jan 13 10:51:27 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Fri Jan 13 10:51:32 2006 Subject: [Visinfo] Syllabus, project info, links to web services apis Message-ID: Hi everyone, The newest syllabus is up, including links to various resources and information about the first project: http://www.mat.ucsb.edu/~g.legrady/academic/courses/06w259/06w259.html ********* Google APIs : http://www.google.com/apis/ http://www.google.com/apis/maps/ Amazon web services pages, including the Alexa search engine web services: http://www.amazon.com/gp/browse.html/ref=sc_fe_l_1/002-7806574-9196022?%5Fencoding=UTF8&node=3435361&no=3435361&me=A36L942TSJ2AJA Yahoo APIs, including Flickr, yahoo maps, yahoo search engine, del.icio.us, etc http://developer.yahoo.net/ Microsoft's MapPoint: http://msdn.microsoft.com/MapPoint/ ESV Bible: http://www.gnpcb.org/esv/share/services/ There are many more out there, If anyone is interested in using web service APIs to gather a dataset, I can help with the REST or SOAP protocols. One interesting idea is to use a "mashup" of different data using various web services APIs. For instance, for the NGDA project I am using Google Maps API in conjunction with the ADL map & air photography api & their gazetteer data . Also, there is a project called Piggybank (http://simile.mit.edu/piggy-bank/) that helps you write screen scrapers for websites that don't expose a programmatic way to access their data. -Angus From angus.forbes at gmail.com Fri Jan 13 11:09:50 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Fri Jan 13 11:09:55 2006 Subject: [Visinfo] Angus' office hours Message-ID: Hi everyone, My office hours are before and after class on Tues or Thurs, where we can meet next door in George's studio. I can also meet on Wednesday after 11pm. -Angus 347-581-1022 On 1/13/06, Jennifer Bernstein wrote: > yup! > hey, when are your office hours? > From jenn_bernstein at yahoo.com Tue Jan 17 11:00:37 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Tue Jan 17 11:00:47 2006 Subject: [Visinfo] Response to Borner Message-ID: <20060117190037.5821.qmail@web81808.mail.mud.yahoo.com> Hi all. I enjoyed this reading. I guess my response is largely questions, along with some big-picture concerns. I?ll start with the nitty gritty. I am still confused about how the clustering techniques relate to the techniques for processing a dataset. As I understand it, you determine frequency and relatedness through an algorithm (SOM, Pathfinder, LSA, etc.). Are the clustering techniques just different ways to determine clusters from those data sets, which in their organic states are simply continuums? I?m not sure I understand the difference between citation linkages and co-occurrence similarities as described on p. 193. Are citation linkages when two documents cite the exact same thing, and co-occurrence similarities when two documents include the same word, but are not referring to the same object as it exists in reality? On page 197, the authors said that when using FA or PCA, you can interpret the factors while you can?t in LSA. Why is that? With LSA, words are determined to be associated if they occur frequently together in a document. Is proximity ever used to determine association? When you use Pathfinder, how do you get the proximity data? Is Pathfinder a secondary technique tagged onto LSA or FA? The authors discuss the deficiencies of the Kohonen algorithm. They say ?These deficiencies comprise the absence of a cost function, and the lack of a theoretical basis for choosing learning rate parameter schedules and neighborhood parameters to ensure topographic ordering.? I don?t understand this statement. Do SOM?s work with smaller datasets? All the above said, my larger concern lies with how semantically legitimate these techniques are. For standardized types of language such as keyword searches, the above seem appropriate. However in content analysis, two sentences could use entirely different words express the same sentiment. Perhaps the similarly would still be caught even if those sentences weren?t associated, but it seems like a lot of meaning and significance could go unnoticed. Or what if two different authors were discussing the same topic with different levels of sophistication? I guess I would like a formal assessment of the type of documents that this word indexing is and isn?t appropriate for. This would help me understand if my results were a factor of the relatedness of the documents or a artifact of the data processing and visualization. responses welcome! -jenn Jennifer Bernstein Masters/Phd Program 4713 Ellison Hall Department of Geography University of California, Santa Barbara Santa Barbara, CA 93101 jennb@geog.ucsb.edu From angus.forbes at gmail.com Wed Jan 18 16:04:30 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Wed Jan 18 16:04:35 2006 Subject: [Visinfo] Andre Skupin lecture Message-ID: Hi class, Here is information about an interesting lecture we'll be required to attend. GPS data and Self-Organizing Maps-- what could be more relevant? -Angus ------------------------ Prof. Andr? Skupin Department of Geography San Diego State University Date: Friday, March 3, 2006 Place: Humanities and Social Sciences, 1173 Time: 2:00 pm ? 3:00 pm (Reception to follow) Abstract: As we move across geographic space, aren't we simultaneously traveling through a high-dimensional attribute space in which the geographic entities are located that we encounter along the way? Of course, such movement may be difficult to imagine in concrete terms, especially when dealing with a very large number of dimensions. In order to aid the human mind in understanding the paths taken during this type of attribute space travel, it is here proposed to create map-like visualizations of high-dimensional paths. A specific methodology is presented for achieving this by integrating a form of artificial neural network known as Kohonen map or self-organizing map (SOM) with space-time paths captured by GPS. Among the envisioned applications are novel forms of surrogate or virtual travel and comparative studies of people's movement across separate geographic territories. A number of case studies serve to illustrate the technique, including a traversal of the Austrian capital, Vienna, and urban travel in the U.S. combined with population attributes for all 200,000+ census block groups. ANDR? SKUPIN is an assistant professor of geography at San Diego State University. He previously held an associate professor position at the University of New Orleans. Areas of interest and expertise include text document visualization, geographic visualization, cartographic generalization, and visual data mining. Much of his research revolves around new perspectives on geographic metaphors, methods, and principles, outside of traditional geographic domains. Recent efforts include the visualization of text documents through a combination of self-organizing neural maps, GIS, and cartographic design. Results of this research have been published within the information science, computer graphics, and cartographic communities, as well as in interdisciplinary outlets, such as the Proceedings of the National Academy of Sciences. Andr? Skupin received a Dipl.-Ing. degree in cartography from the Dresden University of Technology, Germany, and a PhD in geography from the State University of New York at Buffalo. From stacy at geog.ucsb.edu Wed Jan 18 22:23:38 2006 From: stacy at geog.ucsb.edu (Stacy Rebich) Date: Wed Jan 18 22:21:55 2006 Subject: [Visinfo] response to reading Message-ID: <003301c61cc0$e4cf0970$3c6a6f80@phoebe> Well, here it is Response to ?Visualizing Knowledge Domains? by Katy B?rner, Chaomei Chen, and Kevin W. Boyack I enjoyed reading this chapter because it was a good (if relatively superficial) introduction to a variety of data reduction and information visualization techniques that are in use and under development. I think it will be a great resource for future exploration of topics in infoviz since the bibliography is so extensive. I have to admit that since I?m not familiar with many of the statistical/mathematical/programming techniques or fundamentals, the actual concepts that underlie several of the techniques described remain something of a mystery to me. As I was reading the descriptions, I kept thinking, ?it would be really useful to have an illustration for each of these,? and then when I got to the end, I realized that the first endnote provides a link to online illustrations. After looking at these images, the features of the visualizations created through the various approaches became somewhat more apparent. Beyond the specifics of how the individual algorithms work, there is another thing that I?m somewhat confused about. This applies not necessarily only to the techniques described in this chapter, but also to the online tools that I saw when browsing the links on the course website. My question is this: what data type/format is necessary as input type for each method/software package? Which of these approaches require data that has already been organized/reduced in some way, and which are suitable for processing natural language text? I think that in many cases there will first be a categorization step and then a data reduction step, but I?m still unsure about how to choose programs that can help me do each of these. For example, of those that can be used for semantic information visualization, which require that I know the keywords before beginning, and which will find them for me? Which are suitable only for visualizing things that are already discrete entities (authors, papers, etc.)? I didn?t find very specific descriptions of many of the tools anywhere. I guess what I would be interested in seeing is some sort of decision tree or something like that that could help me identify potential viable approach-software combinations for the sorts of analysis/visualization I would like to do. There are two comments in the paper that have helped me to decide on what sort of project I?d like to do for this class. The first is a comment on page 2 of the pdf version: it?s about how one of the problems with traditional approaches to painting the big picture of scientific knowledge is that the survey methods used are subjective. This comment seems to apply that these automated approaches are objective (or at least more objective). Later in the paper, the authors also mention that multiple representations are preferred to achieve a better understanding of the information content. These comments make me think that it would be interesting to implement several of these approaches on a modestly-sized dataset in order to compare the different visualization solutions they arrive at. I?ll talk about this project idea in my presentation. That?s it! I welcome comments, suggestions, or answers to any of my questions. ;) Stacy ~~~~~~~~~~~~~~~~~~ Stacy Rebich Graduate Student Department of Geography University of California Santa Barbara, CA 93106 ~~~~~~~~~~~~~~~~~~ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://zydeco.mat.ucsb.edu/pipermail/visinfo/attachments/20060118/df4f4f36/attachment-0001.html From jenn_bernstein at yahoo.com Thu Jan 19 07:06:16 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Thu Jan 19 07:06:35 2006 Subject: [Visinfo] response to reading In-Reply-To: <003301c61cc0$e4cf0970$3c6a6f80@phoebe> Message-ID: <20060119150618.33935.qmail@web81810.mail.mud.yahoo.com> Hi Stacy, I like the decision tree idea! While the matrix was helpful in terms of the strengths and weaknesses of each type of statistical technique, it wasn't fleshed out enough to be really clear. -jenn --- Stacy Rebich wrote: > Well, here it is > > > > Response to ?Visualizing Knowledge Domains? by Katy > B?rner, Chaomei Chen, > and Kevin W. Boyack > > > > I enjoyed reading this chapter because it was a good > (if relatively > superficial) introduction to a variety of data > reduction and information > visualization techniques that are in use and under > development. I think it > will be a great resource for future exploration of > topics in infoviz since > the bibliography is so extensive. I have to admit > that since I?m not > familiar with many of the > statistical/mathematical/programming techniques or > fundamentals, the actual concepts that underlie > several of the techniques > described remain something of a mystery to me. As I > was reading the > descriptions, I kept thinking, ?it would be really > useful to have an > illustration for each of these,? and then when I got > to the end, I realized > that the first endnote provides a link to online > illustrations. After > looking at these images, the features of the > visualizations created through > the various approaches became somewhat more > apparent. > > > > Beyond the specifics of how the individual > algorithms work, there is another > thing that I?m somewhat confused about. This > applies not necessarily only > to the techniques described in this chapter, but > also to the online tools > that I saw when browsing the links on the course > website. My question is > this: what data type/format is necessary as input > type for each > method/software package? Which of these approaches > require data that has > already been organized/reduced in some way, and > which are suitable for > processing natural language text? I think that in > many cases there will > first be a categorization step and then a data > reduction step, but I?m still > unsure about how to choose programs that can help me > do each of these. For > example, of those that can be used for semantic > information visualization, > which require that I know the keywords before > beginning, and which will find > them for me? Which are suitable only for > visualizing things that are > already discrete entities (authors, papers, etc.)? > I didn?t find very > specific descriptions of many of the tools anywhere. > > > > I guess what I would be interested in seeing is some > sort of decision tree > or something like that that could help me identify > potential viable > approach-software combinations for the sorts of > analysis/visualization I > would like to do. > > > > There are two comments in the paper that have helped > me to decide on what > sort of project I?d like to do for this class. The > first is a comment on > page 2 of the pdf version: it?s about how one of the > problems with > traditional approaches to painting the big picture > of scientific knowledge > is that the survey methods used are subjective. > This comment seems to apply > that these automated approaches are objective (or at > least more objective). > Later in the paper, the authors also mention that > multiple representations > are preferred to achieve a better understanding of > the information content. > These comments make me think that it would be > interesting to implement > several of these approaches on a modestly-sized > dataset in order to compare > the different visualization solutions they arrive > at. I?ll talk about this > project idea in my presentation. > > > > That?s it! I welcome comments, suggestions, or > answers to any of my > questions. ;) > > Stacy > > > > ~~~~~~~~~~~~~~~~~~ > > Stacy Rebich > > Graduate Student > > Department of Geography > > University of California > > Santa Barbara, CA 93106 > > ~~~~~~~~~~~~~~~~~~ > > > > > _______________________________________________ > visinfo mailing list > visinfo@zydeco.mat.ucsb.edu > http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo > From corina1 at umail.ucsb.edu Thu Jan 19 16:42:38 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Thu Jan 19 16:43:41 2006 Subject: [Visinfo] Reading week #1 write up Message-ID: <20060119164238.y6ki6g0cg4kkgwgg@webaccess.umail.ucsb.edu> Corina Schweller MAT259 Reading Week #1 VISUALIZING KNOWLEDGE DOMAINS ? Boerner, Chen, Boyack. The field of Domain Visualization can be very disconnected when viewed from different disciplines. There is a gap between theory and practice, which needs to be bridged. The history of databases, which are often employed for mapping, began in the 1950?s with citation index databases. In the 1960?s mapping was done manually and one of the pioneers was a spatial map of research in DNA. This map allows for scientific communication and analysis of domains. Advances of scientific knowledge can be shown with longitudinal mapping. This type of mapping can even forecast trends. A citation network can be navigated by SCI-Map software, which grows the map based on keywords and is based on clustering. Scientific Visualization is still not very interactive. On the other hand, Information Visualization focuses on interactivity. In the field of geography information can be visualized with geographic coordinates. In order to map information, the corresponding data is necessary. Then the units of analysis need to be selected. The most common units are documents. The Vector Space Model was designed for the retrieval of information. It is utilized for indexing documents and is composed of three parts; document indexing, term weighing, and computing similarity coefficients. The Vector Space Model works according to word matching and allows for a way to find similarities in documents. High dimensional data can be reduced, while still preserving the structure, with techniques such as the Eigenvalue/Eigenvector decomposition. To reduce the number of variables and detect relations of variables the Factor Analysis technique can be employed. The structure between objects in a set of proximity measure can be found with Multidimensional Scaling. Self-Organizing Maps produce a 2D map of the output layer that will show the relationship to the input layer. The Kohonen SOM map algorithm can organize large quantities of information and is used to map the Internet. Information can be organized in various ways. Triangulation maps random points at the origin of a coordinate system . Force Directed Placement sorts randomly placed objects and computes forces between nodes. Semantic Treemaps apply FDP and organize documents via clustering. Visualization can be outlined by the Shneiderman framework; Data Types, Typology of Tasks, Visualizations, and Necessary Features. Fractal Views can visualize large hierarchies and control the amount of information displayed. Less important info is removed and the number of displayed nodes is controlled by fractal dimension. In the future, more robust algorithms are needed to advance information science. More accurate results and a faster response will be the goal of future domain maps. I think mapping has brought much to a visual society and allows us to view data in a more comprehensible manner. The Vector Space Model seems like a clear method of organizing data and retrieval. With these models we can see information displayed according to a method. Mapping shows more than just simple words, it allows us to perceive the similarities and differences between terms with visual spacing and connectivity. From corina1 at umail.ucsb.edu Thu Jan 19 16:43:45 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Thu Jan 19 16:43:53 2006 Subject: [Visinfo] Corina's IGERT write up Message-ID: <20060119164345.5l15y500qoocokgo@webaccess.umail.ucsb.edu> Corina Schweller Week 1 IGERT Ken Goldberg presented the topic of Network Robotics, namely telerobotics, at the Jan 13 IGERT seminar. In 1994 he and a team put a camera on a robot and built a robust system. It was the first network robot on the Internet and allowed people to control certain functions through the website. Users could click on a button that blows a gust of air into a sandbox to uncover items that have been buried in the sand. The next network robot he participated in creating was a robotic arm that allows people to water plants and to plant new ones in a garden via the Internet. At the turn of the century there was an increase for surveillance video for security due to world events. With this shift came a new project at the UC Berkley that employed a powerful camera that could zoom in with incredible clarity. This camera was controlled by a Website interface and utilized algorithms to find the area that would give the most user satisfaction in case there were too many users wanting to change views at the same time. His future project is a Collaborative Observatory for Natural Environments. From godwin at umail.ucsb.edu Sun Jan 22 12:27:37 2006 From: godwin at umail.ucsb.edu (Mike Godwin) Date: Sun Jan 22 12:27:15 2006 Subject: [Visinfo] Godwin on Shedroff's Article Message-ID: <43D3EAB9.2060707@umail.ucsb.edu> An HTML attachment was scrubbed... URL: http://zydeco.mat.ucsb.edu/pipermail/visinfo/attachments/20060122/98b81ec2/attachment.html From godwin at umail.ucsb.edu Sun Jan 22 12:56:15 2006 From: godwin at umail.ucsb.edu (Mike Godwin) Date: Sun Jan 22 12:55:52 2006 Subject: [Visinfo] animation algorithm questions? Message-ID: <43D3F16F.1090605@umail.ucsb.edu> Hey all, Figured I'd throw out some of the questions I've been having with my work, and see if anyone might have answers. Mostly techie Java programming questions, so let me know if this isn't the forum. But I figured we're all going to start struggling with these questions to some degree and maybe there is some collective wisdom out there... 1) Java programming questions. I have an animation and I'm a little unclear about optimum organization of the code -- ie when should one use an array or ArrayList or Vector class? What is the syntax for referring to 2dimensional Vector type arrays? Also I would like input into general object/Class organization - the animated critter is an object, but are each of his legs as well? This will make more sense with a diagram, but any input would be helpful. I'd love a 5 minute chat with a CS inclined individual who might be able to look over some sketches and offer coding structure input. 2) preprocessing / data input. I'll be doing my apple visualizations in processsing, and the files I have are all .dbf. In general I'm trying to figure out how much I should preprocess the data to make it lovely for my visualization program. Seems like a trade-off between extendability (that is if there is a lot of preprocess it'll be harder for me to pull down other datasets and feed them right into the visualization program) and ease of coding. Basically, do I massage my 49 .dbf files into a .csv that would be relatively easy to parse in processing or do battle with the dbf java libraries and thereby automate the massaging in the visualization program. hmm. My new tactic is do whatever = visualization fastest then clean up later, but input on this topic would be helpful as well. cheers, mike From angus.forbes at gmail.com Sun Jan 22 20:52:38 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Mon Jan 23 08:12:32 2006 Subject: [Visinfo] animation algorithm questions? In-Reply-To: <43D3F16F.1090605@umail.ucsb.edu> References: <43D3F16F.1090605@umail.ucsb.edu> Message-ID: hi Mike, Feel free to ask me any questions about Java and/or database connectivity. Answers to those questions depend upon your needs, obviously. But it sounds like your apple visualization will use a small static data set, and so you can probably most easily import into excel, save into a text file, and then read all your data everything directly into memory. Regarding data structures, here's a link to the java collections tutorial: http://java.sun.com/docs/books/tutorial/collections/TOC.html You'll probably want to use the ArrayList implementation of the List interface Bruce Eckel's Thinking In Java is a great resource for thinking about different strategies of modelling objects, etc. It's free from his site: http://www.mindview.net/Books/TIJ/ -Angus On 1/22/06, Mike Godwin wrote: > Hey all, > > Figured I'd throw out some of the questions I've been having with my > work, and see if anyone might have answers. Mostly techie Java > programming questions, so let me know if this isn't the forum. But I > figured we're all going to start struggling with these questions to some > degree and maybe there is some collective wisdom out there... > > 1) Java programming questions. I have an animation and I'm a little > unclear about optimum organization of the code -- ie when should one use > an array or ArrayList or Vector class? What is the syntax for referring > to 2dimensional Vector type arrays? Also I would like input into general > object/Class organization - the animated critter is an object, but are > each of his legs as well? This will make more sense with a diagram, but > any input would be helpful. I'd love a 5 minute chat with a CS inclined > individual who might be able to look over some sketches and offer coding > structure input. > > 2) preprocessing / data input. I'll be doing my apple visualizations in > processsing, and the files I have are all .dbf. In general I'm trying > to figure out how much I should preprocess the data to make it lovely > for my visualization program. Seems like a trade-off between > extendability (that is if there is a lot of preprocess it'll be harder > for me to pull down other datasets and feed them right into the > visualization program) and ease of coding. Basically, do I massage my > 49 .dbf files into a .csv that would be relatively easy to parse in > processing or do battle with the dbf java libraries and thereby automate > the massaging in the visualization program. hmm. My new tactic is do > whatever = visualization fastest then clean up later, but input on this > topic would be helpful as well. > > cheers, mike > _______________________________________________ > visinfo mailing list > visinfo@zydeco.mat.ucsb.edu > http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo > From zmd at umail.ucsb.edu Mon Jan 23 14:51:49 2006 From: zmd at umail.ucsb.edu (Zachary M. Davis) Date: Mon Jan 23 14:51:59 2006 Subject: [Visinfo] Response to 1st Week's Reading Message-ID: <20060123145149.qkmmny95sk84sgwo@webaccess.umail.ucsb.edu> While I believe that this reading will serve as a valuable resource for this course, I feel that its value is confined mainly to the bibliography. It boasts an impressive collection of resources that appear to include much of the important work in the field. However, I felt that the article itself was too superficial to be of any real use. It's obvious from the responses so far that while the article was successful in getting people very superficially acquainted (in the "I've heard of that before" sense) with a wide variety of algorithms, it provided no or little real insight into which methods might be appropriate for a given situation or dataset, and how they differed from eachother. This tendency towards superficial overview, coupled with a tendency to throw out the occasional high-level math concept made the article confusing and difficult to extract any real concepts out of. Also, the authors seemed mostly concerned with visualizations based in citation data, which is all fine and dandy, but doesn't seem to me to warrant the generic title of "Visualizing Knowledge Domains". I didn't dislike the article as much as I fear it's sounding I did, and I certainly have no basis for comparison (although browsing a few other overview papers in the field might be both interesting and worthwhile), I just felt that I didn't gain that much from the reading aside from a sizable (and presumably useful) bibliography. -- Zachary Davis zmd@umail.ucsb.edu From jenn_bernstein at yahoo.com Mon Jan 23 20:02:56 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Mon Jan 23 20:03:06 2006 Subject: [Visinfo] response to Shedroff (week 3) Message-ID: <20060124040256.37220.qmail@web81808.mail.mud.yahoo.com> Hi all, Frankly, I'm not sure when this is due. Here it is anyways. best, -jenn Response to ?Information Interaction Design? by Nathan Shedroff I agree with Mike that this essay seemed both slightly dated and coercive. I imagine that the last 10 years produced more opportunities for a more formalized Information Design education, although perhaps not many. The style reminds me of 50?s advertising, which I find slightly heartwarming. His explanations of information and interaction design could be simplified- information design is visually representing conclusions derived from a dataset, while interaction design allows for exploratory data analysis by the user. The novel part of his essay is the advocacy of sensorial design. It was given a brief treatment in the essay, and I was left wondering why. While I understand the cognitive aspects of how a viewer receives information, I feel like it?s a bit much to take on their sensorial experience as well. One last criticism is the amount of space his graphics occupied. Tufte might take issue with the data/ink ration- you?re not being given a whole lot of information per unit of ink-space. Although I must agree wholeheartedly with Mr. Sheriff when he says ?I believe that one of the nicest experiences you can have is to enjoy a stimulating conversation with another person over great meal.? From jenn_bernstein at yahoo.com Thu Jan 26 10:56:47 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Thu Jan 26 10:56:57 2006 Subject: [Visinfo] Response to InfoVis Cyberinfrastructure Message-ID: <20060126185647.23992.qmail@web81802.mail.mud.yahoo.com> The InfoVis Cyberinfrastructure is a fantastic resource. My only complaint is its incompleteness. I really like the format applied to the data modeling section, where the technique is described, pros and cons are discussed, and then dives in to the nitty-gritty of using the algorithm. It?s too bad that the preprocessing section isn?t done, because I think that?s really where the crux of the issue lies: how do we take meaning and turn it into numbers, both theoretically and practically? Also, when you get down to the data analysis section, it would be nice to have a standardized format to contrast and evaluate the different techniques. Once again, when should and shouldn?t these techniques be used? I like the learning modules, and should probably do a few at some point because I am sure it will contextualize the techniques. -jenn From stacy at geog.ucsb.edu Thu Jan 26 16:40:25 2006 From: stacy at geog.ucsb.edu (Stacy Rebich) Date: Thu Jan 26 16:38:18 2006 Subject: [Visinfo] FW: response to week 2 reading Message-ID: <004401c622da$44791a40$3c6a6f80@phoebe> Hi all, Here's my second attempt to post my response to the IVC site on the listserv. _____ From: Stacy Rebich [mailto:stacy@geog.ucsb.edu] Sent: Wednesday, January 25, 2006 11:44 PM To: 'visinfo-bounces@zydeco.mat.ucsb.edu' Subject: response to week 2 reading Response to Information Visualization CyberInfrastructure website: I think that spending some time going through a lot of the content on this website helped to give me a clearer picture of the general steps I'll need to go through to create information/knowledge from my data. While I can't say I found definitive answers to a lot of my questions here, I do feel that it helped me to develop more specific questions about how to approach my data filtering and organization issues. I'm going to throw in a few of these questions here, and if anyone has some good insight/advice, I'd be happy to hear it. I realize the first step that I need to take once I have my marked-up data is to create a list/matrix of topics that I can use as inputs for the spatialization algorithm(s) I choose. Does anyone have a recommendation for a program that does good stop word removal and stemming? It seems like TMG (see link below) will do these things and construct a matrix.any thoughts about this package? http://scgroup.hpclab.ceid.upatras.gr/scgroup/Projects/TMG/ (Jenn, is this the same one you're using?) I ask about stop word removal and stemming because there is another type of topic extractor I'd like to try as well (described below), but I think it may require some preprocessing of this sort. So the thing I looked at in more detail is the Griffiths and Steyvers Topics Model listed on the IVC page. This model uses latent Dirichlet allocation (see Blei et al. paper). Another application that seems to do basically the same thing is this one at knowceans.org . Anyway, from what I understand of this approach, it's different from other types of topic identification approaches in that it's based on a Bayesian probability model. The algorithm works by first establishing probability estimates for word frequency using a sample set of documents and then uses these probability estimates to identify words/topics in the documents of interest whose frequency exceeds their predicted frequency. It seems that this approach could be a good way to help identify the unique features of each document (uniqueness being determined by the set of sample documents used to establish probability estimates). There was also some discussion in the reading that I did about the need to decide how many topics should be extracted when using this method. It seems that if you ask for a number of unique topics that is too small, the categories turn out to be too general. If you ask for too many, topic groups start to include collections of words that have little obvious connection. Does anyone know if there are any guidelines that can help you to choose a relatively good number of topics to start with? Did anyone else try to download and install the IVC software? I tried, but when I unzipped the download folder, I couldn't find the files discussed in the installation instructions. I think the instructions were actually written for an earlier version, but it wasn't obvious to me how to do the install with the latest release. ~~~~~~~~~~~~~~~~~~ Stacy Rebich Graduate Student Department of Geography University of California Santa Barbara, CA 93106 ~~~~~~~~~~~~~~~~~~ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://zydeco.mat.ucsb.edu/pipermail/visinfo/attachments/20060126/1e3ccf6a/attachment.html From corina1 at umail.ucsb.edu Thu Jan 26 16:39:50 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Thu Jan 26 16:40:00 2006 Subject: [Visinfo] Week 2 reading Message-ID: <20060126163950.w83gy442og4g4kso@webaccess.umail.ucsb.edu> I found the "Graphviz -Graph Visualization Software" to be quite interesting. The data representation is simple, yet efficient and brings a point across with maximum clarity. The Module Dependencies graph has a adequate system of data representation and the layout shows the relationships of the items based on proximity. Another aesthetically pleasing visualization technique is the "Radial Tree". A Radial Tree uses a fisheye technique and has a central node so tha it can display a very large amount of information. Another Software is the The "Fisheye Table" distorts the information so that it can be read more coherently. The Content-Addressable Network reminds me of early math classes and is useful since you can search the graph from node to node. The problem with this system is that it is too centralized. From godwin at umail.ucsb.edu Sun Jan 29 11:12:34 2006 From: godwin at umail.ucsb.edu (Mike Godwin) Date: Sun Jan 29 11:12:12 2006 Subject: [Visinfo] Review of JOONE an alternative Kohonen SOM software pkg Message-ID: <43DD13A2.20202@umail.ucsb.edu> After the SOM_PAK tutorial, I was curious to see if there were any more recent and possibly handier programs for implementing the kohonen algorithm. Angus pointed me towards Joone (Java Object Oriented Neural Engine) so I downloaded it and went through the tutorials... http://www.jooneworld.com/ On the pros: it's java based and therefore easily run from a variety of platforms. It has a very clear and intuitive graphic interface -- one of those drag and drop style programming interfaces, ie. draw a box that represents your input, a box that represents the output, and some boxes that represent your neural network, connect all the arrows and voila! It's really well-suited to visualizing complex neural networks and assessing their learning strengths and faults. Unfortunately Joone is definitely not set up for creative/alternative visualization of the output. It is relatively easy to use the editor to read a table, utilize come customized kohonen algorithm, and then output a table; but it would still be up to you to take that table and parse it into something attractive if data visualization were the ultimate objective. In the end I think this is a wonderful tool for someone more interested in artificial intelligence and neural networks, and aside from browsing the core engine for the implementation of the kohonen algorithm I'll be passing it up for my own work in this class. From legrady at arts.ucsb.edu Sun Jan 29 22:35:10 2006 From: legrady at arts.ucsb.edu (glegrady) Date: Sun Jan 29 22:34:44 2006 Subject: [Visinfo] Interesting Links: In-Reply-To: <004401c622da$44791a40$3c6a6f80@phoebe> References: <004401c622da$44791a40$3c6a6f80@phoebe> Message-ID: <98f123177ff57907161d1b565b8e694d@arts.ucsb.edu> Hi All, Week-end reading: a) http://www.wheresgeorge.com -- tracking dollar bills over time and space and related research: http://www.kitp.ucsb.edu/community/news/hufnagel_jan06.php http://www.kitp.ucsb.edu/community/news/RH_box.jpg http://www.nature.com/nature/journal/v439/n7075/abs/nature04292.html b) Pacific Northwest National Laboratory:Information Visualization (from alex villacorta) http://www.pnl.gov/infoviz/index.html LOts of downloadable papers on InfoVis George Legrady Studio http://www.georgelegrady.com respond to: gl@georgelegrady.com tel.+1.805.637.6195 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 809 bytes Desc: not available Url : http://zydeco.mat.ucsb.edu/pipermail/visinfo/attachments/20060129/c287e2d6/attachment.bin From legrady at arts.ucsb.edu Mon Jan 30 22:40:27 2006 From: legrady at arts.ucsb.edu (glegrady) Date: Mon Jan 30 22:40:12 2006 Subject: [Visinfo] animation algorithm questions? In-Reply-To: <43D3F16F.1090605@umail.ucsb.edu> References: <43D3F16F.1090605@umail.ucsb.edu> Message-ID: <22c4cd8a3846b2b497d9ddbf0200cdb0@arts.ucsb.edu> hi Mike, Not being a computer science expert, I checked with Rama Hoetzlein, my CS trained MAT former TA. Probably Angus would have been able to answer these as well. His answers: The decision to use an array, ArrayList or Vector class really depends on the application. These different structure vary in terms of the way they handle memory. Some do not allow addition or remove (arrays, which are fixed in memory), some allow immediate insertion but slow deletion of objects (linked lists, ie. ArrayList) while others allow slow insertion and deletion but fast searching (Vectors). This site may help you decide: http://www.javaworld.com/javaworld/javaqa/2001-06/03-qa-0622- vector.html If you want to look into these differences further, I would recommend a computer science book on Java data structures. Syntax for 2D vector type arrays: Look it up. There are many Java books at the library. > Also I would like input into general object/class organization - the animated critter is an object, but are each of his legs as well? I presume this is a code design question, ie. how to make an animated critter in which body and legs are both objects/classes. This can be accomplished with classes. An object is an "instance" of a class, just as a cat is an "instance" of an animal (animal describes a whole class of creatures). For example, you might make a "bodyobj" class which also has a reference to the parent object. Then you can make 5 objects, 1 'bodyobj' for the whole critter and 1 for each leg. critter (bodyobj, no parent object) left_front_leg (bodyobj, critter is parent object) right_front_leg (bodyobj, critter is parent object) left_back_leg (bodyobj, critter is parent object) right_back_leg (bodyobj, critter is parent object) Again, a library book on Object-Oriented Java Programming should cover this. Regarding data... > Seems like a trade-off between extendability (that is if there is a lot of preprocess it'll be harder for me to pull down other datasets and feed them right into the visualization program) and ease of coding... This is exactly right, it is a trade off. In general it depends on how flexible you want your input reader to be. If you want to put together something quickly, then figure out the "easiest" way to read data into processing, possibly by reading csv data, and then modify your data using external programs to format it this way (ie. make it csv). This is most often the best choice as it is still fairly flexible (you can just add column/commas). All of the above topics are covered by computer science books that can be found at the library. I don't have specific titles, but here are some search terms: - Data Structure Java (for arrays, vectors and lists) - Object-Oriented Programming Java (for classes and objects) - Database Input Output Java (for data i/o) > On Jan 22, 2006, at 12:56 PM, Mike Godwin wrote: > Hey all, > > Figured I'd throw out some of the questions I've been having with my > work, and see if anyone might have answers. Mostly techie Java > programming questions, so let me know if this isn't the forum. But I > figured we're all going to start struggling with these questions to > some > degree and maybe there is some collective wisdom out there... > > 1) Java programming questions. I have an animation and I'm a little > unclear about optimum organization of the code -- ie when should one > use > an array or ArrayList or Vector class? What is the syntax for referring > to 2dimensional Vector type arrays? Also I would like input into > general > object/Class organization - the animated critter is an object, but are > each of his legs as well? This will make more sense with a diagram, > but > any input would be helpful. I'd love a 5 minute chat with a CS > inclined > individual who might be able to look over some sketches and offer > coding > structure input. > > 2) preprocessing / data input. I'll be doing my apple visualizations > in > processsing, and the files I have are all .dbf. In general I'm trying > to figure out how much I should preprocess the data to make it lovely > for my visualization program. Seems like a trade-off between > extendability (that is if there is a lot of preprocess it'll be harder > for me to pull down other datasets and feed them right into the > visualization program) and ease of coding. Basically, do I massage my > 49 .dbf files into a .csv that would be relatively easy to parse in > processing or do battle with the dbf java libraries and thereby > automate > the massaging in the visualization program. hmm. My new tactic is do > whatever = visualization fastest then clean up later, but input on this > topic would be helpful as well. > > cheers, mike > _______________________________________________ > visinfo mailing list > visinfo@zydeco.mat.ucsb.edu > http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo > > George Legrady University of California, Santa Barbara MAT: http://www.mat.ucsb.edu ART: http://arts.ucsb.edu IGERT: http://media.igert.ucsb.edu STUDIO: http://www.georgelegrady.com tel. 1.805.637.6195 fax. 1.805.563.5752 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 5208 bytes Desc: not available Url : http://zydeco.mat.ucsb.edu/pipermail/visinfo/attachments/20060130/77224907/attachment.bin From zmd at umail.ucsb.edu Tue Jan 31 11:33:52 2006 From: zmd at umail.ucsb.edu (Zachary M. Davis) Date: Tue Jan 31 11:34:02 2006 Subject: [Visinfo] Week 2 response: JUNG Message-ID: <20060131113352.9a6lqh3bvkks08kc@webaccess.umail.ucsb.edu> JUNG is the Java Universal Network/Graph Framework. It is an open source programming library written in Java with the goal of providing a framework for analysis and visualization of data. Because it is derived from the Java API, JUNG shares in some of the inherent benefits of Java, namely portability and superior documentation. JUNG also seems to have an active community behind it, which is alway useful for any programming library. Although I have not had the chance to work with JUNG myself yet, it seems that it is ideal for prototyping visualizations. It provides several data analysis algorithms that could be used for a variety of different datasets. It also provides visualization tools, including, it appears, several ready-made algorithm-visualization combinations as well as the ability to customize. I feel that this framework would be extremely useful for exploring several visualization and/or analysis methods with minimal effort. Once the basic program had been written to get the data and interface the analysis results with the visualization, it seems that it would be fairly trivial to essentially plug in different algorithms or visualizations. I think this would be an excellent prototyping tool for people with at least some basic Java programming knowledge. -- Zachary Davis zmd@umail.ucsb.edu From corina1 at umail.ucsb.edu Tue Jan 31 11:41:59 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Tue Jan 31 11:42:07 2006 Subject: [Visinfo] IGERT week #3 Message-ID: <20060131114159.tqovg65iccgsg4co@webaccess.umail.ucsb.edu> IGERT Week #3 "An Interactional Ethnographic Approach to Video Analysis" Judith Green The IGERT seminar dealt with analyzing and archiving video. Judith Green studies classrooms and public spaces in order to find patterns of action, talk, and activity. She describes the classroom as a 'cultural setting' and archives the human activity that occurs over a period of time. This archiving is related to ethnograpy and the mapping of naturally occurring language and actions. Event mapping represents the flow of conduct between members of a socaila group. It also all depends on the point of view that the event is seen from. Green studies contextualization cues, such as proxemic and kenesic shifts, and the contrast of events in order to map events. Grren used a popular American movie to show how humans change behavior based on evens and interactions. The movie "Groundhog Day" shows how our environment shapes our actions and the effect of humn interaction. From angus.forbes at gmail.com Thu Jan 12 17:09:34 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Thu, 12 Jan 2006 19:09:34 -0600 Subject: [Visinfo] Visinfo mailing list test Message-ID: testing... From nsbe_pr2000 at hotmail.com Thu Jan 12 18:37:51 2006 From: nsbe_pr2000 at hotmail.com (Reginald S. Archer) Date: Thu, 12 Jan 2006 18:37:51 -0800 Subject: [Visinfo] Visinfo mailing list test In-Reply-To: Message-ID: works -----Original Message----- From: visinfo-bounces at zydeco.mat.ucsb.edu [mailto:visinfo-bounces at zydeco.mat.ucsb.edu] On Behalf Of Angus Forbes Sent: Thursday, January 12, 2006 5:10 PM To: visinfo at zydeco.mat.ucsb.edu Subject: [Visinfo] Visinfo mailing list test testing... _______________________________________________ visinfo mailing list visinfo at zydeco.mat.ucsb.edu http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo From angus.forbes at gmail.com Fri Jan 13 10:51:27 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Fri, 13 Jan 2006 12:51:27 -0600 Subject: [Visinfo] Syllabus, project info, links to web services apis Message-ID: Hi everyone, The newest syllabus is up, including links to various resources and information about the first project: http://www.mat.ucsb.edu/~g.legrady/academic/courses/06w259/06w259.html ********* Google APIs : http://www.google.com/apis/ http://www.google.com/apis/maps/ Amazon web services pages, including the Alexa search engine web services: http://www.amazon.com/gp/browse.html/ref=sc_fe_l_1/002-7806574-9196022?%5Fencoding=UTF8&node=3435361&no=3435361&me=A36L942TSJ2AJA Yahoo APIs, including Flickr, yahoo maps, yahoo search engine, del.icio.us, etc http://developer.yahoo.net/ Microsoft's MapPoint: http://msdn.microsoft.com/MapPoint/ ESV Bible: http://www.gnpcb.org/esv/share/services/ There are many more out there, If anyone is interested in using web service APIs to gather a dataset, I can help with the REST or SOAP protocols. One interesting idea is to use a "mashup" of different data using various web services APIs. For instance, for the NGDA project I am using Google Maps API in conjunction with the ADL map & air photography api & their gazetteer data . Also, there is a project called Piggybank (http://simile.mit.edu/piggy-bank/) that helps you write screen scrapers for websites that don't expose a programmatic way to access their data. -Angus From angus.forbes at gmail.com Fri Jan 13 11:09:50 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Fri, 13 Jan 2006 13:09:50 -0600 Subject: [Visinfo] Angus' office hours Message-ID: Hi everyone, My office hours are before and after class on Tues or Thurs, where we can meet next door in George's studio. I can also meet on Wednesday after 11pm. -Angus 347-581-1022 On 1/13/06, Jennifer Bernstein wrote: > yup! > hey, when are your office hours? > From jenn_bernstein at yahoo.com Tue Jan 17 11:00:37 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Tue, 17 Jan 2006 11:00:37 -0800 (PST) Subject: [Visinfo] Response to Borner Message-ID: <20060117190037.5821.qmail@web81808.mail.mud.yahoo.com> Hi all. I enjoyed this reading. I guess my response is largely questions, along with some big-picture concerns. I?ll start with the nitty gritty. I am still confused about how the clustering techniques relate to the techniques for processing a dataset. As I understand it, you determine frequency and relatedness through an algorithm (SOM, Pathfinder, LSA, etc.). Are the clustering techniques just different ways to determine clusters from those data sets, which in their organic states are simply continuums? I?m not sure I understand the difference between citation linkages and co-occurrence similarities as described on p. 193. Are citation linkages when two documents cite the exact same thing, and co-occurrence similarities when two documents include the same word, but are not referring to the same object as it exists in reality? On page 197, the authors said that when using FA or PCA, you can interpret the factors while you can?t in LSA. Why is that? With LSA, words are determined to be associated if they occur frequently together in a document. Is proximity ever used to determine association? When you use Pathfinder, how do you get the proximity data? Is Pathfinder a secondary technique tagged onto LSA or FA? The authors discuss the deficiencies of the Kohonen algorithm. They say ?These deficiencies comprise the absence of a cost function, and the lack of a theoretical basis for choosing learning rate parameter schedules and neighborhood parameters to ensure topographic ordering.? I don?t understand this statement. Do SOM?s work with smaller datasets? All the above said, my larger concern lies with how semantically legitimate these techniques are. For standardized types of language such as keyword searches, the above seem appropriate. However in content analysis, two sentences could use entirely different words express the same sentiment. Perhaps the similarly would still be caught even if those sentences weren?t associated, but it seems like a lot of meaning and significance could go unnoticed. Or what if two different authors were discussing the same topic with different levels of sophistication? I guess I would like a formal assessment of the type of documents that this word indexing is and isn?t appropriate for. This would help me understand if my results were a factor of the relatedness of the documents or a artifact of the data processing and visualization. responses welcome! -jenn Jennifer Bernstein Masters/Phd Program 4713 Ellison Hall Department of Geography University of California, Santa Barbara Santa Barbara, CA 93101 jennb at geog.ucsb.edu From angus.forbes at gmail.com Wed Jan 18 16:04:30 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Wed, 18 Jan 2006 18:04:30 -0600 Subject: [Visinfo] Andre Skupin lecture Message-ID: Hi class, Here is information about an interesting lecture we'll be required to attend. GPS data and Self-Organizing Maps-- what could be more relevant? -Angus ------------------------ Prof. Andr? Skupin Department of Geography San Diego State University Date: Friday, March 3, 2006 Place: Humanities and Social Sciences, 1173 Time: 2:00 pm ? 3:00 pm (Reception to follow) Abstract: As we move across geographic space, aren't we simultaneously traveling through a high-dimensional attribute space in which the geographic entities are located that we encounter along the way? Of course, such movement may be difficult to imagine in concrete terms, especially when dealing with a very large number of dimensions. In order to aid the human mind in understanding the paths taken during this type of attribute space travel, it is here proposed to create map-like visualizations of high-dimensional paths. A specific methodology is presented for achieving this by integrating a form of artificial neural network known as Kohonen map or self-organizing map (SOM) with space-time paths captured by GPS. Among the envisioned applications are novel forms of surrogate or virtual travel and comparative studies of people's movement across separate geographic territories. A number of case studies serve to illustrate the technique, including a traversal of the Austrian capital, Vienna, and urban travel in the U.S. combined with population attributes for all 200,000+ census block groups. ANDR? SKUPIN is an assistant professor of geography at San Diego State University. He previously held an associate professor position at the University of New Orleans. Areas of interest and expertise include text document visualization, geographic visualization, cartographic generalization, and visual data mining. Much of his research revolves around new perspectives on geographic metaphors, methods, and principles, outside of traditional geographic domains. Recent efforts include the visualization of text documents through a combination of self-organizing neural maps, GIS, and cartographic design. Results of this research have been published within the information science, computer graphics, and cartographic communities, as well as in interdisciplinary outlets, such as the Proceedings of the National Academy of Sciences. Andr? Skupin received a Dipl.-Ing. degree in cartography from the Dresden University of Technology, Germany, and a PhD in geography from the State University of New York at Buffalo. From stacy at geog.ucsb.edu Wed Jan 18 22:23:38 2006 From: stacy at geog.ucsb.edu (Stacy Rebich) Date: Wed, 18 Jan 2006 22:23:38 -0800 Subject: [Visinfo] response to reading Message-ID: <003301c61cc0$e4cf0970$3c6a6f80@phoebe> Well, here it is Response to ?Visualizing Knowledge Domains? by Katy B?rner, Chaomei Chen, and Kevin W. Boyack I enjoyed reading this chapter because it was a good (if relatively superficial) introduction to a variety of data reduction and information visualization techniques that are in use and under development. I think it will be a great resource for future exploration of topics in infoviz since the bibliography is so extensive. I have to admit that since I?m not familiar with many of the statistical/mathematical/programming techniques or fundamentals, the actual concepts that underlie several of the techniques described remain something of a mystery to me. As I was reading the descriptions, I kept thinking, ?it would be really useful to have an illustration for each of these,? and then when I got to the end, I realized that the first endnote provides a link to online illustrations. After looking at these images, the features of the visualizations created through the various approaches became somewhat more apparent. Beyond the specifics of how the individual algorithms work, there is another thing that I?m somewhat confused about. This applies not necessarily only to the techniques described in this chapter, but also to the online tools that I saw when browsing the links on the course website. My question is this: what data type/format is necessary as input type for each method/software package? Which of these approaches require data that has already been organized/reduced in some way, and which are suitable for processing natural language text? I think that in many cases there will first be a categorization step and then a data reduction step, but I?m still unsure about how to choose programs that can help me do each of these. For example, of those that can be used for semantic information visualization, which require that I know the keywords before beginning, and which will find them for me? Which are suitable only for visualizing things that are already discrete entities (authors, papers, etc.)? I didn?t find very specific descriptions of many of the tools anywhere. I guess what I would be interested in seeing is some sort of decision tree or something like that that could help me identify potential viable approach-software combinations for the sorts of analysis/visualization I would like to do. There are two comments in the paper that have helped me to decide on what sort of project I?d like to do for this class. The first is a comment on page 2 of the pdf version: it?s about how one of the problems with traditional approaches to painting the big picture of scientific knowledge is that the survey methods used are subjective. This comment seems to apply that these automated approaches are objective (or at least more objective). Later in the paper, the authors also mention that multiple representations are preferred to achieve a better understanding of the information content. These comments make me think that it would be interesting to implement several of these approaches on a modestly-sized dataset in order to compare the different visualization solutions they arrive at. I?ll talk about this project idea in my presentation. That?s it! I welcome comments, suggestions, or answers to any of my questions. ;) Stacy ~~~~~~~~~~~~~~~~~~ Stacy Rebich Graduate Student Department of Geography University of California Santa Barbara, CA 93106 ~~~~~~~~~~~~~~~~~~ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.create.ucsb.edu/pipermail/visinfo/attachments/20060118/df4f4f36/attachment-0002.html From jenn_bernstein at yahoo.com Thu Jan 19 07:06:16 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Thu, 19 Jan 2006 07:06:16 -0800 (PST) Subject: [Visinfo] response to reading In-Reply-To: <003301c61cc0$e4cf0970$3c6a6f80@phoebe> Message-ID: <20060119150618.33935.qmail@web81810.mail.mud.yahoo.com> Hi Stacy, I like the decision tree idea! While the matrix was helpful in terms of the strengths and weaknesses of each type of statistical technique, it wasn't fleshed out enough to be really clear. -jenn --- Stacy Rebich wrote: > Well, here it is > > > > Response to ?Visualizing Knowledge Domains? by Katy > B?rner, Chaomei Chen, > and Kevin W. Boyack > > > > I enjoyed reading this chapter because it was a good > (if relatively > superficial) introduction to a variety of data > reduction and information > visualization techniques that are in use and under > development. I think it > will be a great resource for future exploration of > topics in infoviz since > the bibliography is so extensive. I have to admit > that since I?m not > familiar with many of the > statistical/mathematical/programming techniques or > fundamentals, the actual concepts that underlie > several of the techniques > described remain something of a mystery to me. As I > was reading the > descriptions, I kept thinking, ?it would be really > useful to have an > illustration for each of these,? and then when I got > to the end, I realized > that the first endnote provides a link to online > illustrations. After > looking at these images, the features of the > visualizations created through > the various approaches became somewhat more > apparent. > > > > Beyond the specifics of how the individual > algorithms work, there is another > thing that I?m somewhat confused about. This > applies not necessarily only > to the techniques described in this chapter, but > also to the online tools > that I saw when browsing the links on the course > website. My question is > this: what data type/format is necessary as input > type for each > method/software package? Which of these approaches > require data that has > already been organized/reduced in some way, and > which are suitable for > processing natural language text? I think that in > many cases there will > first be a categorization step and then a data > reduction step, but I?m still > unsure about how to choose programs that can help me > do each of these. For > example, of those that can be used for semantic > information visualization, > which require that I know the keywords before > beginning, and which will find > them for me? Which are suitable only for > visualizing things that are > already discrete entities (authors, papers, etc.)? > I didn?t find very > specific descriptions of many of the tools anywhere. > > > > I guess what I would be interested in seeing is some > sort of decision tree > or something like that that could help me identify > potential viable > approach-software combinations for the sorts of > analysis/visualization I > would like to do. > > > > There are two comments in the paper that have helped > me to decide on what > sort of project I?d like to do for this class. The > first is a comment on > page 2 of the pdf version: it?s about how one of the > problems with > traditional approaches to painting the big picture > of scientific knowledge > is that the survey methods used are subjective. > This comment seems to apply > that these automated approaches are objective (or at > least more objective). > Later in the paper, the authors also mention that > multiple representations > are preferred to achieve a better understanding of > the information content. > These comments make me think that it would be > interesting to implement > several of these approaches on a modestly-sized > dataset in order to compare > the different visualization solutions they arrive > at. I?ll talk about this > project idea in my presentation. > > > > That?s it! I welcome comments, suggestions, or > answers to any of my > questions. ;) > > Stacy > > > > ~~~~~~~~~~~~~~~~~~ > > Stacy Rebich > > Graduate Student > > Department of Geography > > University of California > > Santa Barbara, CA 93106 > > ~~~~~~~~~~~~~~~~~~ > > > > > _______________________________________________ > visinfo mailing list > visinfo at zydeco.mat.ucsb.edu > http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo > From corina1 at umail.ucsb.edu Thu Jan 19 16:42:38 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Thu, 19 Jan 2006 16:42:38 -0800 Subject: [Visinfo] Reading week #1 write up Message-ID: <20060119164238.y6ki6g0cg4kkgwgg@webaccess.umail.ucsb.edu> Corina Schweller MAT259 Reading Week #1 VISUALIZING KNOWLEDGE DOMAINS ? Boerner, Chen, Boyack. The field of Domain Visualization can be very disconnected when viewed from different disciplines. There is a gap between theory and practice, which needs to be bridged. The history of databases, which are often employed for mapping, began in the 1950?s with citation index databases. In the 1960?s mapping was done manually and one of the pioneers was a spatial map of research in DNA. This map allows for scientific communication and analysis of domains. Advances of scientific knowledge can be shown with longitudinal mapping. This type of mapping can even forecast trends. A citation network can be navigated by SCI-Map software, which grows the map based on keywords and is based on clustering. Scientific Visualization is still not very interactive. On the other hand, Information Visualization focuses on interactivity. In the field of geography information can be visualized with geographic coordinates. In order to map information, the corresponding data is necessary. Then the units of analysis need to be selected. The most common units are documents. The Vector Space Model was designed for the retrieval of information. It is utilized for indexing documents and is composed of three parts; document indexing, term weighing, and computing similarity coefficients. The Vector Space Model works according to word matching and allows for a way to find similarities in documents. High dimensional data can be reduced, while still preserving the structure, with techniques such as the Eigenvalue/Eigenvector decomposition. To reduce the number of variables and detect relations of variables the Factor Analysis technique can be employed. The structure between objects in a set of proximity measure can be found with Multidimensional Scaling. Self-Organizing Maps produce a 2D map of the output layer that will show the relationship to the input layer. The Kohonen SOM map algorithm can organize large quantities of information and is used to map the Internet. Information can be organized in various ways. Triangulation maps random points at the origin of a coordinate system . Force Directed Placement sorts randomly placed objects and computes forces between nodes. Semantic Treemaps apply FDP and organize documents via clustering. Visualization can be outlined by the Shneiderman framework; Data Types, Typology of Tasks, Visualizations, and Necessary Features. Fractal Views can visualize large hierarchies and control the amount of information displayed. Less important info is removed and the number of displayed nodes is controlled by fractal dimension. In the future, more robust algorithms are needed to advance information science. More accurate results and a faster response will be the goal of future domain maps. I think mapping has brought much to a visual society and allows us to view data in a more comprehensible manner. The Vector Space Model seems like a clear method of organizing data and retrieval. With these models we can see information displayed according to a method. Mapping shows more than just simple words, it allows us to perceive the similarities and differences between terms with visual spacing and connectivity. From corina1 at umail.ucsb.edu Thu Jan 19 16:43:45 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Thu, 19 Jan 2006 16:43:45 -0800 Subject: [Visinfo] Corina's IGERT write up Message-ID: <20060119164345.5l15y500qoocokgo@webaccess.umail.ucsb.edu> Corina Schweller Week 1 IGERT Ken Goldberg presented the topic of Network Robotics, namely telerobotics, at the Jan 13 IGERT seminar. In 1994 he and a team put a camera on a robot and built a robust system. It was the first network robot on the Internet and allowed people to control certain functions through the website. Users could click on a button that blows a gust of air into a sandbox to uncover items that have been buried in the sand. The next network robot he participated in creating was a robotic arm that allows people to water plants and to plant new ones in a garden via the Internet. At the turn of the century there was an increase for surveillance video for security due to world events. With this shift came a new project at the UC Berkley that employed a powerful camera that could zoom in with incredible clarity. This camera was controlled by a Website interface and utilized algorithms to find the area that would give the most user satisfaction in case there were too many users wanting to change views at the same time. His future project is a Collaborative Observatory for Natural Environments. From godwin at umail.ucsb.edu Sun Jan 22 12:27:37 2006 From: godwin at umail.ucsb.edu (Mike Godwin) Date: Sun, 22 Jan 2006 12:27:37 -0800 Subject: [Visinfo] Godwin on Shedroff's Article Message-ID: <43D3EAB9.2060707@umail.ucsb.edu> An HTML attachment was scrubbed... URL: http://lists.create.ucsb.edu/pipermail/visinfo/attachments/20060122/98b81ec2/attachment-0002.html From godwin at umail.ucsb.edu Sun Jan 22 12:56:15 2006 From: godwin at umail.ucsb.edu (Mike Godwin) Date: Sun, 22 Jan 2006 12:56:15 -0800 Subject: [Visinfo] animation algorithm questions? Message-ID: <43D3F16F.1090605@umail.ucsb.edu> Hey all, Figured I'd throw out some of the questions I've been having with my work, and see if anyone might have answers. Mostly techie Java programming questions, so let me know if this isn't the forum. But I figured we're all going to start struggling with these questions to some degree and maybe there is some collective wisdom out there... 1) Java programming questions. I have an animation and I'm a little unclear about optimum organization of the code -- ie when should one use an array or ArrayList or Vector class? What is the syntax for referring to 2dimensional Vector type arrays? Also I would like input into general object/Class organization - the animated critter is an object, but are each of his legs as well? This will make more sense with a diagram, but any input would be helpful. I'd love a 5 minute chat with a CS inclined individual who might be able to look over some sketches and offer coding structure input. 2) preprocessing / data input. I'll be doing my apple visualizations in processsing, and the files I have are all .dbf. In general I'm trying to figure out how much I should preprocess the data to make it lovely for my visualization program. Seems like a trade-off between extendability (that is if there is a lot of preprocess it'll be harder for me to pull down other datasets and feed them right into the visualization program) and ease of coding. Basically, do I massage my 49 .dbf files into a .csv that would be relatively easy to parse in processing or do battle with the dbf java libraries and thereby automate the massaging in the visualization program. hmm. My new tactic is do whatever = visualization fastest then clean up later, but input on this topic would be helpful as well. cheers, mike From angus.forbes at gmail.com Sun Jan 22 20:52:38 2006 From: angus.forbes at gmail.com (Angus Forbes) Date: Sun, 22 Jan 2006 22:52:38 -0600 Subject: [Visinfo] animation algorithm questions? In-Reply-To: <43D3F16F.1090605@umail.ucsb.edu> References: <43D3F16F.1090605@umail.ucsb.edu> Message-ID: hi Mike, Feel free to ask me any questions about Java and/or database connectivity. Answers to those questions depend upon your needs, obviously. But it sounds like your apple visualization will use a small static data set, and so you can probably most easily import into excel, save into a text file, and then read all your data everything directly into memory. Regarding data structures, here's a link to the java collections tutorial: http://java.sun.com/docs/books/tutorial/collections/TOC.html You'll probably want to use the ArrayList implementation of the List interface Bruce Eckel's Thinking In Java is a great resource for thinking about different strategies of modelling objects, etc. It's free from his site: http://www.mindview.net/Books/TIJ/ -Angus On 1/22/06, Mike Godwin wrote: > Hey all, > > Figured I'd throw out some of the questions I've been having with my > work, and see if anyone might have answers. Mostly techie Java > programming questions, so let me know if this isn't the forum. But I > figured we're all going to start struggling with these questions to some > degree and maybe there is some collective wisdom out there... > > 1) Java programming questions. I have an animation and I'm a little > unclear about optimum organization of the code -- ie when should one use > an array or ArrayList or Vector class? What is the syntax for referring > to 2dimensional Vector type arrays? Also I would like input into general > object/Class organization - the animated critter is an object, but are > each of his legs as well? This will make more sense with a diagram, but > any input would be helpful. I'd love a 5 minute chat with a CS inclined > individual who might be able to look over some sketches and offer coding > structure input. > > 2) preprocessing / data input. I'll be doing my apple visualizations in > processsing, and the files I have are all .dbf. In general I'm trying > to figure out how much I should preprocess the data to make it lovely > for my visualization program. Seems like a trade-off between > extendability (that is if there is a lot of preprocess it'll be harder > for me to pull down other datasets and feed them right into the > visualization program) and ease of coding. Basically, do I massage my > 49 .dbf files into a .csv that would be relatively easy to parse in > processing or do battle with the dbf java libraries and thereby automate > the massaging in the visualization program. hmm. My new tactic is do > whatever = visualization fastest then clean up later, but input on this > topic would be helpful as well. > > cheers, mike > _______________________________________________ > visinfo mailing list > visinfo at zydeco.mat.ucsb.edu > http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo > From zmd at umail.ucsb.edu Mon Jan 23 14:51:49 2006 From: zmd at umail.ucsb.edu (Zachary M. Davis) Date: Mon, 23 Jan 2006 14:51:49 -0800 Subject: [Visinfo] Response to 1st Week's Reading Message-ID: <20060123145149.qkmmny95sk84sgwo@webaccess.umail.ucsb.edu> While I believe that this reading will serve as a valuable resource for this course, I feel that its value is confined mainly to the bibliography. It boasts an impressive collection of resources that appear to include much of the important work in the field. However, I felt that the article itself was too superficial to be of any real use. It's obvious from the responses so far that while the article was successful in getting people very superficially acquainted (in the "I've heard of that before" sense) with a wide variety of algorithms, it provided no or little real insight into which methods might be appropriate for a given situation or dataset, and how they differed from eachother. This tendency towards superficial overview, coupled with a tendency to throw out the occasional high-level math concept made the article confusing and difficult to extract any real concepts out of. Also, the authors seemed mostly concerned with visualizations based in citation data, which is all fine and dandy, but doesn't seem to me to warrant the generic title of "Visualizing Knowledge Domains". I didn't dislike the article as much as I fear it's sounding I did, and I certainly have no basis for comparison (although browsing a few other overview papers in the field might be both interesting and worthwhile), I just felt that I didn't gain that much from the reading aside from a sizable (and presumably useful) bibliography. -- Zachary Davis zmd at umail.ucsb.edu From jenn_bernstein at yahoo.com Mon Jan 23 20:02:56 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Mon, 23 Jan 2006 20:02:56 -0800 (PST) Subject: [Visinfo] response to Shedroff (week 3) Message-ID: <20060124040256.37220.qmail@web81808.mail.mud.yahoo.com> Hi all, Frankly, I'm not sure when this is due. Here it is anyways. best, -jenn Response to ?Information Interaction Design? by Nathan Shedroff I agree with Mike that this essay seemed both slightly dated and coercive. I imagine that the last 10 years produced more opportunities for a more formalized Information Design education, although perhaps not many. The style reminds me of 50?s advertising, which I find slightly heartwarming. His explanations of information and interaction design could be simplified- information design is visually representing conclusions derived from a dataset, while interaction design allows for exploratory data analysis by the user. The novel part of his essay is the advocacy of sensorial design. It was given a brief treatment in the essay, and I was left wondering why. While I understand the cognitive aspects of how a viewer receives information, I feel like it?s a bit much to take on their sensorial experience as well. One last criticism is the amount of space his graphics occupied. Tufte might take issue with the data/ink ration- you?re not being given a whole lot of information per unit of ink-space. Although I must agree wholeheartedly with Mr. Sheriff when he says ?I believe that one of the nicest experiences you can have is to enjoy a stimulating conversation with another person over great meal.? From jenn_bernstein at yahoo.com Thu Jan 26 10:56:47 2006 From: jenn_bernstein at yahoo.com (Jennifer Bernstein) Date: Thu, 26 Jan 2006 10:56:47 -0800 (PST) Subject: [Visinfo] Response to InfoVis Cyberinfrastructure Message-ID: <20060126185647.23992.qmail@web81802.mail.mud.yahoo.com> The InfoVis Cyberinfrastructure is a fantastic resource. My only complaint is its incompleteness. I really like the format applied to the data modeling section, where the technique is described, pros and cons are discussed, and then dives in to the nitty-gritty of using the algorithm. It?s too bad that the preprocessing section isn?t done, because I think that?s really where the crux of the issue lies: how do we take meaning and turn it into numbers, both theoretically and practically? Also, when you get down to the data analysis section, it would be nice to have a standardized format to contrast and evaluate the different techniques. Once again, when should and shouldn?t these techniques be used? I like the learning modules, and should probably do a few at some point because I am sure it will contextualize the techniques. -jenn From stacy at geog.ucsb.edu Thu Jan 26 16:40:25 2006 From: stacy at geog.ucsb.edu (Stacy Rebich) Date: Thu, 26 Jan 2006 16:40:25 -0800 Subject: [Visinfo] FW: response to week 2 reading Message-ID: <004401c622da$44791a40$3c6a6f80@phoebe> Hi all, Here's my second attempt to post my response to the IVC site on the listserv. _____ From: Stacy Rebich [mailto:stacy at geog.ucsb.edu] Sent: Wednesday, January 25, 2006 11:44 PM To: 'visinfo-bounces at zydeco.mat.ucsb.edu' Subject: response to week 2 reading Response to Information Visualization CyberInfrastructure website: I think that spending some time going through a lot of the content on this website helped to give me a clearer picture of the general steps I'll need to go through to create information/knowledge from my data. While I can't say I found definitive answers to a lot of my questions here, I do feel that it helped me to develop more specific questions about how to approach my data filtering and organization issues. I'm going to throw in a few of these questions here, and if anyone has some good insight/advice, I'd be happy to hear it. I realize the first step that I need to take once I have my marked-up data is to create a list/matrix of topics that I can use as inputs for the spatialization algorithm(s) I choose. Does anyone have a recommendation for a program that does good stop word removal and stemming? It seems like TMG (see link below) will do these things and construct a matrix.any thoughts about this package? http://scgroup.hpclab.ceid.upatras.gr/scgroup/Projects/TMG/ (Jenn, is this the same one you're using?) I ask about stop word removal and stemming because there is another type of topic extractor I'd like to try as well (described below), but I think it may require some preprocessing of this sort. So the thing I looked at in more detail is the Griffiths and Steyvers Topics Model listed on the IVC page. This model uses latent Dirichlet allocation (see Blei et al. paper). Another application that seems to do basically the same thing is this one at knowceans.org . Anyway, from what I understand of this approach, it's different from other types of topic identification approaches in that it's based on a Bayesian probability model. The algorithm works by first establishing probability estimates for word frequency using a sample set of documents and then uses these probability estimates to identify words/topics in the documents of interest whose frequency exceeds their predicted frequency. It seems that this approach could be a good way to help identify the unique features of each document (uniqueness being determined by the set of sample documents used to establish probability estimates). There was also some discussion in the reading that I did about the need to decide how many topics should be extracted when using this method. It seems that if you ask for a number of unique topics that is too small, the categories turn out to be too general. If you ask for too many, topic groups start to include collections of words that have little obvious connection. Does anyone know if there are any guidelines that can help you to choose a relatively good number of topics to start with? Did anyone else try to download and install the IVC software? I tried, but when I unzipped the download folder, I couldn't find the files discussed in the installation instructions. I think the instructions were actually written for an earlier version, but it wasn't obvious to me how to do the install with the latest release. ~~~~~~~~~~~~~~~~~~ Stacy Rebich Graduate Student Department of Geography University of California Santa Barbara, CA 93106 ~~~~~~~~~~~~~~~~~~ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.create.ucsb.edu/pipermail/visinfo/attachments/20060126/1e3ccf6a/attachment-0002.html From corina1 at umail.ucsb.edu Thu Jan 26 16:39:50 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Thu, 26 Jan 2006 16:39:50 -0800 Subject: [Visinfo] Week 2 reading Message-ID: <20060126163950.w83gy442og4g4kso@webaccess.umail.ucsb.edu> I found the "Graphviz -Graph Visualization Software" to be quite interesting. The data representation is simple, yet efficient and brings a point across with maximum clarity. The Module Dependencies graph has a adequate system of data representation and the layout shows the relationships of the items based on proximity. Another aesthetically pleasing visualization technique is the "Radial Tree". A Radial Tree uses a fisheye technique and has a central node so tha it can display a very large amount of information. Another Software is the The "Fisheye Table" distorts the information so that it can be read more coherently. The Content-Addressable Network reminds me of early math classes and is useful since you can search the graph from node to node. The problem with this system is that it is too centralized. From godwin at umail.ucsb.edu Sun Jan 29 11:12:34 2006 From: godwin at umail.ucsb.edu (Mike Godwin) Date: Sun, 29 Jan 2006 11:12:34 -0800 Subject: [Visinfo] Review of JOONE an alternative Kohonen SOM software pkg Message-ID: <43DD13A2.20202@umail.ucsb.edu> After the SOM_PAK tutorial, I was curious to see if there were any more recent and possibly handier programs for implementing the kohonen algorithm. Angus pointed me towards Joone (Java Object Oriented Neural Engine) so I downloaded it and went through the tutorials... http://www.jooneworld.com/ On the pros: it's java based and therefore easily run from a variety of platforms. It has a very clear and intuitive graphic interface -- one of those drag and drop style programming interfaces, ie. draw a box that represents your input, a box that represents the output, and some boxes that represent your neural network, connect all the arrows and voila! It's really well-suited to visualizing complex neural networks and assessing their learning strengths and faults. Unfortunately Joone is definitely not set up for creative/alternative visualization of the output. It is relatively easy to use the editor to read a table, utilize come customized kohonen algorithm, and then output a table; but it would still be up to you to take that table and parse it into something attractive if data visualization were the ultimate objective. In the end I think this is a wonderful tool for someone more interested in artificial intelligence and neural networks, and aside from browsing the core engine for the implementation of the kohonen algorithm I'll be passing it up for my own work in this class. From legrady at arts.ucsb.edu Sun Jan 29 22:35:10 2006 From: legrady at arts.ucsb.edu (glegrady) Date: Sun, 29 Jan 2006 22:35:10 -0800 Subject: [Visinfo] Interesting Links: In-Reply-To: <004401c622da$44791a40$3c6a6f80@phoebe> References: <004401c622da$44791a40$3c6a6f80@phoebe> Message-ID: <98f123177ff57907161d1b565b8e694d@arts.ucsb.edu> Hi All, Week-end reading: a) http://www.wheresgeorge.com -- tracking dollar bills over time and space and related research: http://www.kitp.ucsb.edu/community/news/hufnagel_jan06.php http://www.kitp.ucsb.edu/community/news/RH_box.jpg http://www.nature.com/nature/journal/v439/n7075/abs/nature04292.html b) Pacific Northwest National Laboratory:Information Visualization (from alex villacorta) http://www.pnl.gov/infoviz/index.html LOts of downloadable papers on InfoVis George Legrady Studio http://www.georgelegrady.com respond to: gl at georgelegrady.com tel.+1.805.637.6195 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 809 bytes Desc: not available Url : http://lists.create.ucsb.edu/pipermail/visinfo/attachments/20060129/c287e2d6/attachment-0002.bin From legrady at arts.ucsb.edu Mon Jan 30 22:40:27 2006 From: legrady at arts.ucsb.edu (glegrady) Date: Mon, 30 Jan 2006 22:40:27 -0800 Subject: [Visinfo] animation algorithm questions? In-Reply-To: <43D3F16F.1090605@umail.ucsb.edu> References: <43D3F16F.1090605@umail.ucsb.edu> Message-ID: <22c4cd8a3846b2b497d9ddbf0200cdb0@arts.ucsb.edu> hi Mike, Not being a computer science expert, I checked with Rama Hoetzlein, my CS trained MAT former TA. Probably Angus would have been able to answer these as well. His answers: The decision to use an array, ArrayList or Vector class really depends on the application. These different structure vary in terms of the way they handle memory. Some do not allow addition or remove (arrays, which are fixed in memory), some allow immediate insertion but slow deletion of objects (linked lists, ie. ArrayList) while others allow slow insertion and deletion but fast searching (Vectors). This site may help you decide: http://www.javaworld.com/javaworld/javaqa/2001-06/03-qa-0622- vector.html If you want to look into these differences further, I would recommend a computer science book on Java data structures. Syntax for 2D vector type arrays: Look it up. There are many Java books at the library. > Also I would like input into general object/class organization - the animated critter is an object, but are each of his legs as well? I presume this is a code design question, ie. how to make an animated critter in which body and legs are both objects/classes. This can be accomplished with classes. An object is an "instance" of a class, just as a cat is an "instance" of an animal (animal describes a whole class of creatures). For example, you might make a "bodyobj" class which also has a reference to the parent object. Then you can make 5 objects, 1 'bodyobj' for the whole critter and 1 for each leg. critter (bodyobj, no parent object) left_front_leg (bodyobj, critter is parent object) right_front_leg (bodyobj, critter is parent object) left_back_leg (bodyobj, critter is parent object) right_back_leg (bodyobj, critter is parent object) Again, a library book on Object-Oriented Java Programming should cover this. Regarding data... > Seems like a trade-off between extendability (that is if there is a lot of preprocess it'll be harder for me to pull down other datasets and feed them right into the visualization program) and ease of coding... This is exactly right, it is a trade off. In general it depends on how flexible you want your input reader to be. If you want to put together something quickly, then figure out the "easiest" way to read data into processing, possibly by reading csv data, and then modify your data using external programs to format it this way (ie. make it csv). This is most often the best choice as it is still fairly flexible (you can just add column/commas). All of the above topics are covered by computer science books that can be found at the library. I don't have specific titles, but here are some search terms: - Data Structure Java (for arrays, vectors and lists) - Object-Oriented Programming Java (for classes and objects) - Database Input Output Java (for data i/o) > On Jan 22, 2006, at 12:56 PM, Mike Godwin wrote: > Hey all, > > Figured I'd throw out some of the questions I've been having with my > work, and see if anyone might have answers. Mostly techie Java > programming questions, so let me know if this isn't the forum. But I > figured we're all going to start struggling with these questions to > some > degree and maybe there is some collective wisdom out there... > > 1) Java programming questions. I have an animation and I'm a little > unclear about optimum organization of the code -- ie when should one > use > an array or ArrayList or Vector class? What is the syntax for referring > to 2dimensional Vector type arrays? Also I would like input into > general > object/Class organization - the animated critter is an object, but are > each of his legs as well? This will make more sense with a diagram, > but > any input would be helpful. I'd love a 5 minute chat with a CS > inclined > individual who might be able to look over some sketches and offer > coding > structure input. > > 2) preprocessing / data input. I'll be doing my apple visualizations > in > processsing, and the files I have are all .dbf. In general I'm trying > to figure out how much I should preprocess the data to make it lovely > for my visualization program. Seems like a trade-off between > extendability (that is if there is a lot of preprocess it'll be harder > for me to pull down other datasets and feed them right into the > visualization program) and ease of coding. Basically, do I massage my > 49 .dbf files into a .csv that would be relatively easy to parse in > processing or do battle with the dbf java libraries and thereby > automate > the massaging in the visualization program. hmm. My new tactic is do > whatever = visualization fastest then clean up later, but input on this > topic would be helpful as well. > > cheers, mike > _______________________________________________ > visinfo mailing list > visinfo at zydeco.mat.ucsb.edu > http://zydeco.mat.ucsb.edu/mailman/listinfo/visinfo > > George Legrady University of California, Santa Barbara MAT: http://www.mat.ucsb.edu ART: http://arts.ucsb.edu IGERT: http://media.igert.ucsb.edu STUDIO: http://www.georgelegrady.com tel. 1.805.637.6195 fax. 1.805.563.5752 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 5208 bytes Desc: not available Url : http://lists.create.ucsb.edu/pipermail/visinfo/attachments/20060130/77224907/attachment-0002.bin From zmd at umail.ucsb.edu Tue Jan 31 11:33:52 2006 From: zmd at umail.ucsb.edu (Zachary M. Davis) Date: Tue, 31 Jan 2006 11:33:52 -0800 Subject: [Visinfo] Week 2 response: JUNG Message-ID: <20060131113352.9a6lqh3bvkks08kc@webaccess.umail.ucsb.edu> JUNG is the Java Universal Network/Graph Framework. It is an open source programming library written in Java with the goal of providing a framework for analysis and visualization of data. Because it is derived from the Java API, JUNG shares in some of the inherent benefits of Java, namely portability and superior documentation. JUNG also seems to have an active community behind it, which is alway useful for any programming library. Although I have not had the chance to work with JUNG myself yet, it seems that it is ideal for prototyping visualizations. It provides several data analysis algorithms that could be used for a variety of different datasets. It also provides visualization tools, including, it appears, several ready-made algorithm-visualization combinations as well as the ability to customize. I feel that this framework would be extremely useful for exploring several visualization and/or analysis methods with minimal effort. Once the basic program had been written to get the data and interface the analysis results with the visualization, it seems that it would be fairly trivial to essentially plug in different algorithms or visualizations. I think this would be an excellent prototyping tool for people with at least some basic Java programming knowledge. -- Zachary Davis zmd at umail.ucsb.edu From corina1 at umail.ucsb.edu Tue Jan 31 11:41:59 2006 From: corina1 at umail.ucsb.edu (Corina S. Schweller) Date: Tue, 31 Jan 2006 11:41:59 -0800 Subject: [Visinfo] IGERT week #3 Message-ID: <20060131114159.tqovg65iccgsg4co@webaccess.umail.ucsb.edu> IGERT Week #3 "An Interactional Ethnographic Approach to Video Analysis" Judith Green The IGERT seminar dealt with analyzing and archiving video. Judith Green studies classrooms and public spaces in order to find patterns of action, talk, and activity. She describes the classroom as a 'cultural setting' and archives the human activity that occurs over a period of time. This archiving is related to ethnograpy and the mapping of naturally occurring language and actions. Event mapping represents the flow of conduct between members of a socaila group. It also all depends on the point of view that the event is seen from. Green studies contextualization cues, such as proxemic and kenesic shifts, and the contrast of events in order to map events. Grren used a popular American movie to show how humans change behavior based on evens and interactions. The movie "Groundhog Day" shows how our environment shapes our actions and the effect of humn interaction.