Wednesday, September 21, 2011

Reaction: The Eyes Have It: A Task By Data Type Taxonomy For Information Visualization

This paper has been instrumental in forming a foundation and creating some ground rules and methods to study and create visualizations. The paper divides the types of data that is generally visualized into different categories. It categorizes the data based on characteristics of the data and also based on the way such data needs to be visualized. For example temporal data is separate category instead of being in 2D data. The author has explained the categories very well and given a definition for each category. He has also provided examples for each category.

The author has formulated the basic mantra of visualization. He has suggested what kind of visualizations can be used for each category and how the basic tasks are fulfilled by such a visualization. It was a very neat paper to read and provided a very good foundation for data visualization.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The paper starts of highlighting the fact that information visualization as a field focuses on presentation of the visualization and has very less focus on interaction. In modern interfaces even the basic tasks on visualizations have interactions. So this brings up a valid point. The author talks about different interaction styles derived from  human computer interaction like direct manipulation. In order to address this problem the author has surveyed a lot of visualizations and recorded all the basic types of interactions a user does on visualization. The survey has resulted in seven categories of interaction. He has then talked about these categories. The list seems exhaustive for the present day interfaces. Although this might change with changes in systems of interactions and new interaction devices.

I felt that the paper has succeeded in bringing out the point about interaction. It has also provided a starting point to devise interaction strategies.  It was a good read and different kind of paper.

Reaction: A Review of Overview+Detail, Zooming, and Focusing+Context Interfaces

This paper analyzes tasks overview+detail, zooming, and focusing, cue based interfaces. There is a good amount of analyses of the presentation and interaction techniques for these tasks. The paper is full of examples and detailed illustrations which explain which interaction techniques work and which dont. There is a good amount of reasoning provided too for these examples. The examples using google maps was really good as I could relate to most of the points mentioned. Using commonly used applications for criticizing is one of the good things about this paper as it makes the ideas more concrete.  The paper seems to be lengthy and wordy. The paper not only talks about good interaction techniques but also criticized some of the techniques. 

Viz:, the Graphical Resume imports your Linkedin information and generates it as a timeline-esque infographic. The visualization is created with HTML and SVG. Other templates and options are available to tweak and customize the visualization. The creator is also looking to open up a "theme marketplace" to allow other developers and designers to create new templates.

I wonder if this will actually become a common method of advertising yourself to potential employers.

via TechCrunch: "Graphical Resume Site Vizualize.Me Launches, We Talk To The Founder."

Reaction:Toward a Deeper Understanding of the Role of Interaction in Information Visualization

In this paper as the author suggests we get to read information on the role of interaction in Information visualization. A branch of web visualization is human interaction with interface.Hence we understand the importance of interaction wrt information visualization.

while designing an interface one should always keep in mind the user. How the intended user wishes to interact with the interface is a key understanding. This paper tells us how and what user looks for specifically wrt to interaction. There various categories of interaction that the author discusses in this paper.

It is necessary for interface design engineers to understand the method of interaction the user looks for. Hence this paper provides the approach needed for the above information.

Reaction:A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

In this paper the author puts light on the numerous interface strategies. As the topic suggests the topics discussed are overview+detail,zooming and focus+context.

The paper is a good example for web visualization and interaction of users with user interfaces. Various approaches that can be taken to improve the above by using individually or combining the strategies the author has made aware of in the paper.

Various illustrations have been provided to help in better understanding of the approaches. By reading this paper due to the shortcomings and benefits mentioned on each approach enough undersatnding can be gained for future research.

Announcement: privacy and FERPA forms


Most of you have not turned in your FERPA forms, so let me take this opportunity to stress privacy on our various course sites.

In the broadest spirit of learning, I prefer to make our course content as accessible to the public as possible.

However, to protect your privacy, our blog and forum must be private (accessible only to us) until all of you return your signed forms, or tell me that you don't plan to. Once I have heard from all of you, I will remove any content from those who prefer to remain private, and make the sites public. In this way we can both protect your privacy and reach a broader audience, including future students. Similarly, you will not be allowed to add content to the public wiki until you sign the form.

Keep in mind that you are not required to sign the FERPA form, which essentially says that you're okay with anyone seeing what you add and knowing you're in this class (we will never reveal grades and feedback). You can simply email me any future comments, helpful suggestions, or reactions. If you would prefer not to make your projects public, you can email us a zip file for viewing locally.

I would ask, however, that as soon as possible, all of you either return the form (email or hardcopy), or let me know via email that you would prefer not to.



Reaction : Toward a Deeper Understanding of the Role of Interaction in Information Visualization

This paper focuses on two important facets of Information Visualization: "representation" and "interaction". The author feels that interaction has always been on the back-burner and the contemporary practices give more importance to representation.Thus via this paper, the author tries to convey the advantages we can achieve by focusing on the "interaction" aspect of InfoViz.

The paper makes a comparison saying that interaction methods like zoom, select, filter etc are as important as the selection of charts and colors to be used. Focusing on interaction, a designer comes to know about the end-user's intent. This knowledge comes in handy while visualizing information.

This paper has given me a realization that effective tools are more powerful than pretty ones. The main aim for information visualization is easy interpretations of data and better interactivity helps the purpose.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

This paper covers UI designing strategies of a visualization tool.Main strategies include: overview +Detail, zooming, context+focus, cue. In a real life project these strategies come handy while designing the interface for a tool. This paper provides a detailed road-map for the same. Further in the paper, the author has examples to illustrate that there can not be a specific guideline to distinguish interface. For ex: Zoom is same as to focus + context.

There are lots of examples which make the paper look good initially. However without a few implementation details, I lost interest in it. There is a comprehensive research and a lot of coverage on relevant topics in this paper. According to me this is both the selling point and downside of this paper. As the author tries to cover a lot but misses a clear central idea for the paper.

Reaction: The Eyes have it: A Task by Data Type Taxonomy for Information Visualizations

This paper categorizes information visualizations into seven data types. The paper includes seven tasks that can be performed on the visualization to get the most information out of them.

This paper is based on the central idea that to analyze any information visualization we should first get an overview of the entire set of data. Further filter out the data which we are not interested and finally get the details on demand.

I found this paper interesting. It gave me an idea of the order of information retrieval. It also focused the relevant viewing types. These were more meaningful because of the accompanying examples.

Tuesday, September 20, 2011

Visualizaton: Visualize Me is a web service that imports your Linkedin resume information and expresses it as an infographic. It’s a neat idea and an appropriate format for the attention-span-challenged medium in which it lives…the Internet.

The goal, as CEO Eugene Woo puts it, is to “reinvent the resume by building something more relevant, more visual and more dynamic. One way to do that is by transforming your text resume into an infographic”.

Reaction: A Review of Overview+Detail, Zooming, and Focusing+Context Interfaces

This paper includes a plethora of details regarding interface design techniques, of which discussed are the overview+detail, zooming, context+focus, and cue. As far as project design goes, this is the most useful so far in planning appropriate interface strategies. It has, as the article states, provided this road map of interface construction using the relevant techniques discussed.  In  reading this article, I was struck by the breadth and depth of information discussed. This extended to each of the categories of interface discussed, and will help provide a foundation for, and understanding of, the particular interface we decide upon for our projects. Even with all the attached additional imagery showing particular examples, with so many additional examples discussed, they could have either discussed a few less examples, or provided a little more information on implementation of the examples that they did provide images for.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

This paper provides what seems to be a very good overview of techniques for constructing a useful and informative visualization. This, in fact, has provide something of a template for constructing the visualization for our class projects. It was particularly useful in providing a categorical view of selection techniques, comparing and contrasting these techniques to provide a stronger indication of the usefulness for a specific application. What would have been more useful would be to provide, in addition to descriptions of these categories and visualizations of their applicability, more in-depth coverage of how to implement these techniques. Though this is an overview article, this could have been integrated into each categorical subsection.

Reaction: The Eyes Have It: A Task By Data Type Taxonomy For Information Visualization

Like the previous paper “Toward a Deeper Understanding of the Role of Interaction in Information Visualization”, this paper provides strong direction in generating visualizations; however, as opposed to the review of categorical techniques for visual interactions, this paper provides an ordering to information access which categorical techniques can be used to access. This paper was especially helpful in not only providing this order of information retrieval, but also in elucidating the relevant viewing types and linking these to specific examples which take advantage of this ordering. Providing direct visual examples would have made it easier to work with the relevant examples as opposed to simply referencing them.

Announcement: university web servers for your projects


I've spoken with Carlos in our department's IT service about hosting for your projects and received this reply. My only question is that I think engineering may remove these sites once you graduate, but they may do for now at least. Let me know if you have any questions:

Let me suggest Personal Web sites through Engineering Web hosting services. See for more information. That environment should be adequate for javascript and HTML5 the only caveat is that this service is for engineering students only. Although the disk quota limit is 2GB, given the nature of the project, I think that amount should be sufficient. Let me also mention that students can request (optional) a database along with their personal wewb locker; however, the database is secured so that only connections from their personal web slace is accepted. If you have students go with the above solution, I suggest they request those lockers at least a week in advance. For non-enginerring students the university also offers personal web page space please see 

Either solution above can be accessed remotely via ssh, sftp.


Monday, September 19, 2011

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The paper describes the various interface strategies that can be used for designing the user interface of a visualization tool. The approaches that are discussed are overview +Detail, zooming, context+focus, cue.

According to me, the paper has a huge amount of information. The author has put in a lot of work which is clear from the extensive amount of illustrations that has been provided for each interface strategy. There can never be a one unique user interface for a particular visualization tool. Each type of user interface strategy has its pros and cons. The author has even illustrated with examples that there may not always be a clear distinction between the type of user interface. For ex: Zooming strategy, in a way, is similar to focus + context.The fish eye effect an the hyperbolic tree browser were quite an interesting read.

Reaction: The Eyes have it: A Task by Data Type Taxonomy for Information Visualizations

This paper presents an interesting fundamental approach in seeking information from advance graphical interfaces and visualizations.

According to the author, information visualizations can be categorized into seven data types. He also defines seven tasks that can be performed on the visualization to get the most information out of them.

The author insists that to analyze any information visualization we should first get an overview of the entire set of data. Once we have the overview, we should filter out the data which we are not interested in. And the last step is to get the details on demand. I think we can relate this approach to plain data as well. For example, if someone is searching statistical data about city he would filter out the locality, region etc first.

I particularly liked the explanation where author corelates the dynamic queries with the direct-manipulation queries that are used for advance filtering. With regard to graphic visualizations, this is quite critical as with the emergence of Web 3.0 (HTML5 and CSS3), it may be difficult to quickly fetch structured data and display it fast enough on the user interfaces. As an explanation, it's surely better to invert the dataset while using NOT operator rather than fetching the entire set of data from the database.

Overall, I think this is an interesting paper that deals with future prospects of efficiently analyzing ever increasing information using graphical visualizations.

Reaction: Toward a deeper Understanding of the Role of Interaction in Information Visualization

The author starts off by describing the relationship between interaction and infoviz. He has given two definitions for interaction: “the communication between user and the system” and “Direct manipulation and instantaneous change”. Both these definitions simply mean how the user handles the system provided to him!

Then the author talks about different categories of interaction. The main question asked here is what the user wants to achieve through his interaction with the visualization. I think it’s very important to closely examine the needs of different users before deciding on the interaction styles of your infoviz. Here the author talks about seven different categories of interactions and all of them according to me are very important. The user should be able to select, explore-look for more, reconfigure, encode, abstract, filter and connect. I don’t think any of the interaction styles were new to me but having read about then will really help me in designing my own visualization as I will know what users generally look for when they deal with an infoviz.

Reaction : A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

This was an informative paper to read. Web visualization is a part of Human Computer Interaction and thus there has to be a lot of emphsis laid on the interaction part. In my opinion this paper is dedicated to that part. It provides detailed approaches that one could follow while designing an interface for user interaction. The authors have organized the approaches into four categories which distinguish systems according to their varying uses of space, time, or visual effect. The title itself makes it clear what approaches are presented in the paper. The authors have tried to state, whenever possible, which features can be combined together. For examle- zooming and overview+ detail could be combined. But they have avoided detailed discussion on such combinations to focus on isolated features. The supporting diagrams or examples for almost every presented concept make this paper much easier and better to understand. What I liked the most about this paper is the in depth coverage of each approach including illustrations, usages, pros and cons. Still the length of the paper make it bit difficult to grasp.

Reaction: The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

In this paper the author has described how should one go about while planning for making a graphical interface design or visualization. The author insist on following few steps like first you should carry out the overview, then zoom and filter and later on focus on the details-on-demand. What he is trying to say here is that people/designers should gain an overview of all entire data available in the collection, the remove the data that you feel is not really interesting or unimportant and select an item and get details when needed. The he describes the different types of data that is available now a days like 1-dimensional, 2-dimensional, 3-dimensional, multidimensional , temporal, tree, network etc.

The author is absolutely right in saying that visualizing data definitely gives the user a better understanding of the information. The user finds it relatively easier to deal with an image or a graphical representation than scrolling through large sheets of data. He explains this by giving an example of the world map. If the user knows in which country a particular city is located then it becomes very easy for him to locate it on the map rather than scrolling through the sheet of data alphabetically. The design can always be made better by taking the advantage of the human perceptual abilities.

Reaction : Toward a Deeper Understanding of the Role of Interaction in Information Visualization

Author point out the two core components of Infovis: "representation" and "interactiion". He argues that the focus has always been on the former and interaction is often not giving the same amount of time and research as representation. This paper mainly focuses on the "iteraction" aspect of infovis , by first identifying the fundamental ways in which the users can interact and benefit from it. Then looking back into the work that has already been done in this regard and provide analysis of the way in which research is done with respect to "interaction". The author comes with an notion of category based on the user intent. "select","explore" etc which are taken into detail and explained with an example and a picture. Also one of the examples shown in the class "visual thesarus" is explained in "Explore" category. The categories serve as a good classification , however which might not be adequate. The fact that interaction is as important as representation is the most important thing to take away from this paper.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

This paper reviews the four different approaches namely overview+detail, Zooming and panning, focusing and context and cue-techniques; towards designing interfaces for the users to use in any information visualization applications. The author puts forth the mechanisms in which the approaches put forth the items in the information space. The paper further discusses the advantageous properties and the pitfalls of the four approaches.

I like the way the author has discussed each approach's basics and how it carries out the interaction with the user. Further more, the author provides examples of applications in which the approach has been used and mentions how it is beneficial and how sometimes the interaction confuses the user. I particularly liked the Google Maps Zoom functionality example in which users do not know how to zoom out by double right-clicking. I could relate this to myself!

Overall, it is a good but a very lengthy paper. The information value is really good and interesting.  The paper puts forth subtle pitfalls of the four approaches and also mentions how they can be overcome it.

Reaction: A Review of Overview+ Detail, Zooming, and Focus+ Context Interfaces

This is a very interesting paper that deals with different aspects of visualization like focussing, zooming, overview+detail display, focus+context interfaces, panning etc. This paper basically summarizes the state of research on interfaces that allow users to work at multiple levels of detail and to identify effective and ineffective uses of them.

The author has used several examples in this paper to put across his point. The most interesting among these were the Document Lens which uses continuous functions to diminish document regions with distance from the focus and the mixed resolution large display using the 1024*768 LCD which is an excellent example for panning. In this the images are stitched together in software to ensure that panning actions in one image causes the changes to take place in the corresponding image as well. I also like the idea of the hyperbolic tree browser and Fisheve distortion effects with icon-panels. I think it is an interesting concept.

Reaction: The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

The paper considers several data types and tasks that can be applied on them for maximum cognition. When we consider a type of data, it becomes necessary to understand the attributes and dimensionality of the data at hand. This helps to choose the most appropriate task to be performed on the data.

Several important characteristics of each of the seven data types are considered with examples in the real world for each of these types. This makes it easier for the reader to understand the characteristics of the data. The tasks are then discussed to analyze which task is more suitable for that data set. The paper offers examples and reasons which make it an interesting and informative read.

Reaction : A review of overview+ detail, zooming, and focus+ context interfaces

One interesting thing about Overview+detail is that several applications allow thumbnails sizes to be changed according to the viewer’s choice. This reduces trade-offs between time to access data and visual clarity. But most users are unaware of such features in an application. It was also evident that use of lenses is more suitable towards application such as "draw" and less suitable to text viewing applications.

We see several practical issues mentioned in the paper in our day to day lives. For example, the paper mentions that users are unaware of how to zoom (using mouse click) or how to reverse it. In an IPod map view, a single finger tap zooms in and a double finger tap zooms out. People generally know how to zoom in but not how to zoom out.

Further, Focus+context techniques in source code navigation have really made the lives of programmers easy!

Reaction: Toward A Deeper Understanding of the Role of Interaction in Information Visualization

The paper talks about the importance of Interaction in Information Visualization. One interesting thing the paper mentions is that even with a static image, a user performs several interactions. The paper identifies several categories of interactions in InfoVis based on what the user wants to achieve from an interaction. This "user intent" criteria sounded appropriate and most prominent.

The author lists several categories of interaction and provides examples to portray the use of the method. This makes the method easier to understand in context. Among the categories, Reconfigure sounded important and interesting due to the application of jitter, in which the position of items is shifted to uncover many more items in a region.

Reaction: Toward A Deeper Understanding of the Role of Interaction in Information Visualization

The paper evaluated seven different catagories of interaction that are widely used today. Initially the paper describe InfoVis as a combination of visualization and interaction. Throughtout the paper, the author attempts to reinforce the idea that interation is equally important as it can reveal hidden characteristics of data and make the relationship between data known to the user.

Overall, after reading the paper, I am convinced that while attractive visualization may grab the attention of the user, interaction proves to be a critical piece to having very useful visualization. Afterall, the point of InfoVis is to reduce the amount of work needed to understand data, and in combination with good visualiztion, good interaction would do just that.

Reaction: A Review Of Overview+Detail, Zooming and Focus+Context Interfaces

The paper did a very thorough job of explaining different approaches to interfaces, their advantages and disadvantages. The author provided good real-life examples of each of the four interface approaches, which help the reader corrolate to its arguments much more easily. As it turns out, no one approach stood out to be the best. Rather certain approach fit better than others when it's used for a specific purpose.

I'll be interesting in seeing more of these interface approaches to make it into the apps/OS we use daily. Additionally, if we could investigate whether mixing some of the techniques used by multiple approaches is a good idea and if it produce an interface that contains the best of both worlds.

Reaction: The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

This paper overviews seven high level tasks needed to explore information effectively in software. By separating out different dimension of data, the author clearly identify how to tackle each task that correspond to that dimension. In addition, filtering techniques could be used to narrow down specific information. However the author did describe that the problem with English, unlike programming languages, is that it can't concisely define what the user is looking for. While this may be the case, perhaps some advance form of tagging may alleviate the problem. Overall, this paper present some good approaches to querying and gathering information using the seven tasks.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The author of the paper starts by pointing the fact that interactions make visualizations more useful. This is especially important in the present day. Increasingly data visualization is being used as one of the ways to explore data and also to find serendipitous patterns. A lot of emphasis has been given to rendering a good visualization. However there is still a lot of work that needs to be done in the area of interactions with visualization. These two topics are studied extensively in isolation and need to be combined in a harmonious way to provide more effective visualizations. 

The author takes the first steps by categorizing the types of interactions with visualizations seen so far into seven categories. The author lists and describes these categories. The author provides examples for all these seven basic types of interactions. He also provides examples where the interactions techniques intersect or are incomplete and how this affects the interaction. In conclusion paper shows that there needs to be good amount of focus on the user intent too while making a visualization. 

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

This article focusses more on the interface (system) aspect of information visualization that improves the user interaction.

I feel that this paper exposes the different interfaces to the core illustrating examples and showing how really things work. This paper is excellent for people who want to understand the finer details and not just the abstraction or theory. The article has made compelling attempts to explain techniques through scientific theory, illustrations, giving real time examples. Above all, I enjoyed reading FOCUS+CONTEXT section because it allows to show the universe, yet focus the object currently under the lens. The DataLens fisheye calender tool is one which I found to be more interesting.

The types are evaluated based on performance, usability, cognition. The paper concludes with giving us a tradeoff between the different interface types and the situations in which they might come handy and situations where they are not.

I think that the length of the paper is justified after understanding the various interface aspects it covers.

Reaction : A review of overview+ detail, zooming, and focus+ context interfaces

The author mainly focuses on the fact that everything that we interact with the computer is through a defined screen , which many have scrolling or windows which are mostly accepted standard interfaces. Still they introduce a discontinuity between the information displayed at different times and places. The main focus is to have a summary of interfaces that allow users to work at different levels and with a degree of detail which is effective to them. The "overview+details","focus+details" & "fisheye views" interface is very well explained by giving the Google Maps example. One of the good points of this paper is the snapshots of the exact problems that the author is talking about , like the thumbnails which we can clearly see the difference of having 5 versus 10 in the giving canvas.Later on there is also "cue based" , so all in all the 4 interfaces are dealt in detail and also supported with empirical study. Each of the interfaces have strengths however overview+detail and zooming interfaces have now kind of become the standard for most the desktops.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The paper is a very descriptive. It has a detailed survey of techniques used in Overview+Detail, Zooming, and Focus+Context Interfaces. The paper has rich examples with diagrams. It does a good job in bringing out the different ways the basic visualization tasks are implemented. It talks about how these tasks can be implemented and what are some of the practices that work. It provides reasoning for techniques that work using granular tasks and also cognitive loads. Unlike many other papers this survey evaluates the examples against the verified results in the field. The Mac OS dock fish eye example is interesting. The discussion on google maps gives some idea about zooming and how it needs to be really good or it might hinder the over all use of the application itself.

The paper has been summarized very well. It first enlists all the approaches and describes them in simple terms. It also points out that there is no ideal way of building a visualization and approaches and tasks to be used depend on the objective the visualization. 

Reactions: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

We had read in one of the previous papers that a picture is more than a thousand words. This paper laid more emphasis on making that picture more interactive than just presenting a static image. When it comes to visualization of data, a static image is not much better than a table of data. Both are not very easy to comprehend. I totally agree with the paper that to provide an effective visualization to the end-user there are two main components, namely- representation of data using graphics and the interaction between the user & the visualized data. Providing basic interaction methods like zoom, select, filter etc is as important as the selection of charts and colors to be used. Also I like the section when the authors talk about the user intent. The end-user intent should be the primary focus while designing a visualization. The classification of the interaction techniques, which could be applied at various levels of granularity, is the heart of this paper. The ultimate goal is to provide the user with interaction techniques that give then the ability to manipulate the representations directly or indirectly and hence interpret it more easily. This paper has surely broadened my perspective.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The paper gives us a deep insight into the importance of interaction in information visualization. The author stresses the need of interaction in any form of visualization by categorizing the different ways in which it is actually used implicitly. Although researchers have worked on the representation part of the data sets in information visualization, there has been less work on the interaction part of the information visualization, and the author quotes this by referring to the recent publications in the field of information visualization.

The author stresses the importance of interaction by making a point that even static images can be considered to have an implicit way of communicating information from the system to the user through passive interaction. The author talks about different taxonomies that were considered for understanding Information visualization and stresses on the importance of having considering interaction to be taken more seriously in trying to build better systems. The paper gives us an insight into differentiating the users of the model to the viewers of the model.

The author also gives us their approach in coming with a solution for building better models by making it very clear that interaction plays a significant role in info-viz systems. The way in which they have elucidated their work giving us an overview of the efforts really made me help get ideas on building up for my work in providing me direction.

The paper helps in categorizing different techniques that can be used to discuss interaction techniques. The paper provides for giving out information on the scope for research in information visualization.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interface

This is an excellent review paper to help people wishing to work on similar fields to gain knowledge of what interface mechanisms were available till the date of the paper. It being a paper of the past four years it is pretty much useful in understanding the current state-of-the-art in the field of data visualization particularly concentrating in interface strategies.

The author tries to categorize the interface mechanisms to four, Overview + Detail, Zooming and Focus+Context and by doing so, successfully allows users to attain both focussed and contextual views of the data samples. The author tries to separate the data samples to be viewed either as a focussed view or a contextual view to be able to comprehend the complete information that is meant to be and according to him, any system that manages to do so is close to represent an ideal one. Most of the research work that has happened post this paper have results that resemble what this paper has to say on a high-level abstraction and is well supported by the studies on similar grounds. This paper is an exhaustive information list for the ways interface mechanisms are to be built and is an easy paper to read with lots of information.

Reaction: The eyes have it: a task by data type taxonomy for information visualizations

This paper which dates back to July of 1996 is one of those papers written when visualization was still a nascent and unexplored area. This paper actually gives us a clear understanding of why there was a need of shift towards visualization of information from normal data retrieval systems. The author stresses on the point, "Overview first, zoom and filter then details on demand", which was the need of an hour when data retrieval systems were just not successful in handling large systems as the users intended them to be.

The authors have come up with one of the better taxonomies among their contemporary researchers by diving the tasks and data types into categories of simple and more pragmatic in dealing with the problems then faced. The seven tasks and the data types represent a high level abstraction of the needs of having visualization to handle large data sets. Although this was not the only paper that talks about task by data type taxonomy, this paper successfully defines and covers all aspects that are to be considered while taken all attributes into the system.

Although this paper is written way back in the late 1990's it gives a great insight into how the then researchers had to look at the early stages of visualization and the needs of importance of catering to the users/viewers have enabled them to the unexplored areas of visualization.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

This is very interesting, comprehensive and systematic article which focuses on categorizing the interaction techniques, which is one of the two main components of visualization. The author thinks that little attention is drawn to area of Interactions compared to its "big brother" representation. The author feels a need to connect user objectives with interaction techniques, which have been researched separately so far.

The author presents seven categories of interactions. The nomenclature used is appealing with each technique precisely named using one-two words. The categories are explained clearly with good examples and images. I found Encode and Filter as the most interesting categories.

The methods section of the article is interesting read which describes how authors focus shifted from types of visualization to intent of user while categorizing the interaction techniques. The last part of the article truly contributes to the completeness of the paper. The author gives examples of incompleteness and intersections of interaction categories discussed above. He also explains robustness and usefulness of user-intent-centric categorization. The articles excellent read and attempts to direct visualization designers to think from the users' perspective and intentions while designing visualization.

Reaction: The eyes have it: A task by data type Taxonomy for Information Visualizations

Ben Shneiderman's paper lays out very well the role of information visualization in a world where information overload is rampant. The paper points out the the simple guidelines for a good visualization in a single sentence. "Over first, zoom and filter, then details on demand". Even today these are the tasks that are of prime importance in a visualization. The author tries to classify the tasks based on data type taxonomy. The taxonomy suggested by the author takes into consideration the common types of data being visualized. There are details about the how they are different and why some of them like temporal data and network data deserve special category. Then the paper lists the important ways of zoom in and filter for each type of data type. The paper also lists the different techniques that have been used to perform the basic tasks on different data types. It is in way a state of the art for information visualization in 1996. 

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The paper discusses three visualization schemes Overview+Detail, Zooming and Focus+Context. The paper gives detailed explanation of each category with example and figures. Further the author refers to some empirical evaluation studies to compare the effectiveness of each category. The authors infers few important points like, Overview+Detail is better for understanding of text based documents whereas, focus+context technique can provide better usability in terms of cosmetic effects. Overview+Detail and zooming along with panning usually act as complementary. 

The author warns us that incorrectly implemented zooming techniques can be difficult for user to understand and can hamper effectiveness of zooming. The article shows how focus+context can be helpful for non-spatial representation. The working of Mac OS X Dock is interesting read. The article also mentions cue based techniques which can help to focus on information of interest. Overall I found this article very useful and informative for visualization schemes.

Reaction: The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

This is a basic article which proposes a 'type by task taxonomy (TTT)' of infoviz. Here the author categorizes the visualizations in seven data types and defines seven tasks to be performed on the visualizations. The author underlines the Visual Information Seeking mantra to be 'Overview first, zoom and filter, then details-on-demand'.

The author has described the data types at a very introductory level without much in-depth discussion. The examples given are good, but inclusion of some images and figures could have made the understanding easier. The article explains tasks like Overview, zoom and filter that are basic but probably the most important. The article also draws attention to consider the tasks of history and extracts as important part of Information visualization.

The end of the article however, seems to lose the track and fails to reach a firm conclusion. That part could have been better utilized to reinforce the relationship between the data types and tasks.

Reaction : The Eyes Have it

The author mainly starts with the problems of representing huge amounts of data(collections). For example it is easy to represent a information from a page but if we have hundreds of books , it is a difficult task. Also the color visualization of the huge amounts of data is the area of research. He advocates his mantra of Visual Information by simple : Overview first , zoom and filter then details on demand. He then proposes the task by data type taxonomy and identifies 7 data types with a detailed section on each( 1-d,2-d ,temporal,network,multidimensional etc). The paper is well composed with articulate examples. The idea of simple overview, zoom and if necessary give details along with a defined structure of his data types is a good approach. However there are hurdles of software compatibility and user training and most importantly adoption of this universally to have a same user experience.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The article takes inspirations from one of the recent books on interaction and makes an effort to support the stand that interaction needs attention and research than its sibling-representation. I feel this article and the "The Eyes Have It" article go hand in hand and stress that the interaction component of information visuation given relatively more importance.

It gives an account of the different taxonomies of interaction techniques which touch different aspects and areas of interaction from the system and the user point of view.

The authors have done a comprehensive analysis of past papers and commercial articles to select different interaction techniques. They found that grouping the articles by "user intent" is more pratical and sound than any other classification.
Again, the paper coincides with the idea of providing abstract details and then drilling down to the finer details and filtering out unwanted details, reconfiguring. All these methods help to understand what the user is trying to achieve at each point by going to and fro in the representation.

In its discussion section, the article exposes some interaction techniques which cannot be classified fully because of their ability to fulfill multiple user intents in one go. The article is based on categorization of the different interaction techniques to understand them extensively based on common and uncommon features.

I found the article to be really informative than the 'Eyes have it' article even though they relate to the same idea and concepts. This article I find has more finer details and the area where this article wins over is that in its very outlook is user centric. The article acknowledges that this paper is one of the initial steps in establiship interaction as a full fledged science in itself.

Reactions: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

This paper talks about the four approaches which are overview+detail, zooming, focus+context, and cue-based techniques. The objective of this article is to summarize the state of research on interfaces that allow users to work at multiple levels of detail, and to identify effective and ineffective uses of them. The main aim is to provide a brief summary of the state-of-the-art in interfaces for working with focused and contextual views, and in the understanding of their performance issues as derived from empirical studies.

One good thing about the paper is that it has given diagrams for each and every example like FishEye and the document lens, which make understand the concept better.

Low-Level evaluations of mechanical manipulation and target acquisition and High-Level evaluations including comprehension of information spaces are a very clear and concise review of Overview+Detail, Zooming, and Focus+Context Interfaces.

Overall the paper is informative but I found it a little too descriptive. Even the summary is of two pages.

Reaction: The Eyes Have It - A Task by Data Type Taxonomy for Information Visualizations

What a paper to This paper could truely be used as a guideline for doing stuff related to information visualization. They have broadly covered most of the data types and interaction methods one could use while visualizing the information. There is a difference between just presenting a static visual image and providing an interactive visualization. This papers does tell us about putting more stress on a better data visualization. Also I totally agree with the fact that providing more features and interaction to the user might be overwhelming initially but once the user has practiced and learned it becomes really easy to comprehend and more informative. When I think as a beginner I found the seven tasks to cover most methods of interaction. Zoom, unzoom, filter, navigate, etc are all really basic but very important tasks. Overall this paper has succeeded in changing my percpective about visualizing data and has given me some new ideas which we could include in the project.

Reactions: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The paper seeks to identify the fundamental ways that interaction is used in Infovis systems and its benefits. The authors have explained the paper in a very systematic manner. First they have signified the importance of interaction in Infovis systems and revealed its subtle complexity then provided a novel user intent-based categorization to discuss and characterize interaction techniques in Infovis ultimately showing why interaction is so necessary for Infovis systems.

Few of the points that authors have put up to corroborate their points:

  • Infovis technique or system becomes a static image or autonomously animated images without interaction techniques. Operations such as moving a dynamic query slider to narrow the set of data points being shown or selecting an alternate point in a fisheye view to change the focus seem like clear examples of interactive behavior.
  • Interaction techniques in Infovis seem more designed for changing and adjusting visual representation than for entering data into systems, which clearly is an important aspect of interaction in HCI.
  • Interaction techniques in Infovis are features that provide users with the ability to directly or indirectly manipulate and interpret representations.
  • By supporting further exploration of data items, interaction enables users to have multiple perspectives and gain insight on the data set. It is what separates an Infovis system from a static image. They conclude that these two components are in a symbiotic relationship.

The paper lists taxonomies in tabular form and infers that they lack something important. The first three sets focus strongly on interaction techniques and are relatively system-centric. The last set focuses on user goals without a main focus on interaction. I completely agree with authors that it would be beneficial to bridge these two efforts to connect user objectives with the interaction techniques that help accomplish them.

Reaction: Eyes have it

The author starts off by speaking of a mantra for designing advanced graphical user interfaces i.e. the visual information seeking Mantra: overview first, zoom and filter, then details on demand. Author very well states that as computer speed and display resolution increase, information visualization and graphical interfaces are likely to have an broadened role and proposes a task by data type taxonomy with seven data types and seven tasks.

I like the way he explains all the seven data types and tasks by giving previous, current and future relevant and informative examples like for 3-dimensional data type the challenge offered is to navigate the images of human body in National Library of Medicine's Visible Human Project. Another examples include the technique of parallel coordinates in multi dimensional data type and Fisheye strategy for overview of the entire collection.

I agree with the author's point that the striking perceptual abilities of humans are not at all utilized in current user interfaces and hence if explored can give rise to many new remarkable opportunities. According to the author the unique information exploration tools like dynamic queries, tree maps, fisheye views, parallel coordinated, star fields, and perspective walls are few inventions that will we have to validated and studied.

I even agree that for commercial products to succeed they will have to accommodate not just one but many dimensional datas and provide smooth integration with existing software and support the full task list.

Reaction:A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

    The paper is too lengthy. I was not able to read it through completely.The authors have discussed about  different types of interfaces. The paper reviews each type and the combination of different techniques used for visualization. I found the examples to be interesting and also the fact that discontinuity  should be minimum.The paper talks about challenges in adopting different combo of techniques.I The fish eye effect in Mac OS  proves that a technique should be used only when it improves cognition.

Reaction: A review of overview+detail, zooming, and focus+context interfaces

This article takes some of the interfaces described in "The Eyes Have It" and pairs them together to form dual interfaces that allow users to see details about large sets of data by manipulating the interface. The discussion about the human eyes in this article is perplexing. It is interesting that our viewing angle without head movement is incredibly large and most displays fit within this area, but it is still difficult to produce visualizations that allow the viewer to easily comprehend complex data.

The limitation of our viewing interpretations put into perspective the importance of interaction techniques suggested for implementation by "The Eyes Have It" and "Toward a Deeper Understanding of the Role of Interaction in Information Visualization." The authors also add that sometimes interaction techniques are used for applications that do not need them. I agree that interaction techniques should only be added if they improve the interface or visualization's functionality. Their example of the Mac OS X dock is a proper example to critique. While the "fish-eye" effect is visually appealing, it provides no additional functionality. In my experience, running into the issue of missing the target on the dock is not uncommon, but the visual appeal motivates me to keep the interaction technique enabled.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The authors state that one of the motives behind the paper is the inclusion of advanced interface developing techniques in desktops. I strongly consider this to be an important one because not all interfaces developed by developers have favorable positive results. Thus research helps them understand this difference of perception and ultimately leads in developing better products.

The authors base their ideas on the "mantras" that are described in the paper "The Eyes Have It - A Task by Data Type Taxonomy for Information Visualizations". I feel that by doing so, the authors have presented a step by step gathering of ideas which translate into clear understanding from the readers point of view. Examples of each interface are clearly presented with graphical images.

The paper is an enjoyable read because it provides guidelines which are to be followed when dealing with interfaces and it also describes the situations which might arise when we try to combine multiple views. Some of the views presented in the paper have limited applications or can be easily substituted with others which can serve a basis for viewing multiple representations. I am impressed about the facts of the human eye and its viewing capability . I feel the authors correctly conveyed the fact that adding interfaces which serve limited purposes can only increase the complexity while being of little purpose. Apart from the MAC OSX docking example that is given, we can recollect several instances where introducing a new interface did prove to be of much use.

I also liked the way "Summary" was described by the author. It provides a complete gist of the entire paper and with the use of sub-headings in bold, it mentions key notes to be taken away from the paper.

Reaction: The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

The author points out that the digital libraries that will be developed in the near future will enable convenient exploration of data. He also argues that with increased processing and capacity, visualizations will have a greater role. They can help in many tasks if not all to convey the data, patterns hidden inside the data, statistics. The author attributes it to the human visual bandwidth is greater than any of the other senses and this enables us to grasp data more effectively.

I agree with this point that the author makes because its a common experience especially with huge statistical data like server utilizations or metrics, that one would like to have data displayed as charts and graphs than text. I feel visualizations become efficient and usable when they provide an abstrated view of the data and then as the user demands for finer details, they expose the underlying details.

The article interests me the most in section 3 where it has introduced the seven data types and their tasks. It almost gives me an impression that the author has created a JAVA class of visualizations having seven data types and the following tasks on them. These tasks, I feel, are realistic and handy especially for large data sets where there is a lot of analysis required.

On the whole, the paper lays the foundation about various categories of visualizations and the ideal tasks that must be supported.

Reaction:A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The paper focus on understanding different inteface mechanisms that helps users with effective communication of information. With the assumption that advanced interaction techniques are already in place and there is need for evaluating their efficiency, authors review four categories based on usage of space,time and visual effects. Spatial Separation for simultaneous display of overview and detailed view can be achieved using scroll bars thumbnails etc, applications of which can be seen in google maps and Microsoft PowerPoint. Authors advocate one-way synchronization between overview and detailed view as slight decoupling allow users to access overview without altering detailed view. Design issues associated with zooming are well supported with examples like lack of awareness about activation and deactivation of zooms and space-scale diagram to show increase in amount of panning with increase in scale and desert fog. I enjoyed reading several Automatic zooming theories like depth modulated flying, SDAZ, orthoZoom controller. Focus+Context approach largely rely on fish-eye concept but can also achieved by regulating size and resolution of display. Mac-OS display panel and Microsoft word views effectively convey gist of Fisheye concept. Fourth approach, Cue-based, can be used in conjunction with any of the above approaches.Its widely used as an effective search technique and for conveying out-of-scope information to users.

Empirical analysis and categorical review of diffrent theories based on low-level and high level evaluations clearly explain advantage of using one approach over another as well as combination of approaches.Conclusion drawn include, Fisheye performs better that the overview+details and zooming while dealing with of target acquisition and multiple foci. Overview+details outperforms zooming techniques in terms of task completion times for tasks. However, non-overview is better than overview + details, in general. For the purpose of comprehension, overview+details better than fisheye as it promotes focused reading.Overall its a very informative read . Every approach is backed up with related theories and experimental proof.

Reaction: The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

This paper is the starting point for anyone planning to design advanced GUIs. The author states that design can be bettered by taking advantage of the underutilized human perceptual abilities. This paper stresses on the visualization mantra of "overview first, zoom, filter and details on demand". The author explains the different tasks that a user has to perform to actually decipher data if it falls in one of the seven categories mentioned in the paper. The examples which are stated helped me visualize how a user can interpret the data.
For a user to analyze the data, the visualization should provide him with the basic tasks. This relation has been neatly put forth in the paper. On the other hand, the data types mentioned are not the only ones in existence and many more will emerge as the amount of data grows. These subsequently would require more complex tasks for analysis.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

According to author for complete end-user experience in InfoViz systems is driven by two main components:- representation of data using graphics and interaction with end users. Author argues that shifting focus to interaction, which has been overshadowed by representation, can actually overcome limitations of representaion and can improve performance of infoviz systems as a whole.The goal of the paper is to develop a taxonomy of interaction techniques which can be applied at different level of granularity. Authors studied diffrent taxonomies and arrived at a convincing definition of interaction techniques as the features that provide users with the ability to directly or indirectly manipulate and interpret representations.However that does not take into account static images. Affinity diagramming method helped in concluding that representation and interaction techniques are strongly coupled and hence user intent is an appropriate parameter to classify interaction techniques.Further,from their studies, author came up with seven categories of interaction which can lead to efficient and systematic data exploration.Author explains every category with supporting theories and examples like Direct walk is explained with reference to visual thesaurus for exploring data;Reconfiguration examples include SDM and Conetress, Attribute explorer support encoding, also be achieved by color-encoding, drill-down operation in a tree map visualization to depict abstraction is explained through SequoiaView, filtering can be done using dynamic query controls and author also illustrate connect interactions with the help of vizster and name voyager. Overall, its a very informative read and leaves room for improvement by stating "categories are not collectively exhaustive".

Reaction : The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

According to author, while dealing with large volume of information,data warehousing and visualization takes precedence over database maintenance. Information communicated visually is most effective. Its easier to perceive data graphically than textually. The type by task taxonomy serve as a guide for representaing multiattribute data in seven diffrent ways and performing seven diffrent tasks on these data types to solve the issues in focus. Author follows a mantra as a part of visual design guidelines :- "Overview first, Zoom and filter,then details on demand". Since these data types and tasks represent a higher-level of abstraction,they can be used in combination and hold room for further addition and refinement. Filtration of uninteresting items, maintenance of history for record-keeping and rollback , and extraction adds a lot of flexiblity to the existing process.The importance of human perceptual ability for visual information is justified by author as "The eye, the hand, and the mind seem to work smoothly and rapidly as users perform actions on visual displays." Overall its an informative read. I particularly liked reading about the filter flow model for direct manipulation queries that cleanly deals with ambiguities in boolean expressions and diffrent ways of representing indented data like tree-structure and networks.

Reaction:Toward a Deeper Understanding of the Role of Interaction in Information Visualization

        This paper focuses on the interaction component of Info viz. It claims that research is focused on representation part but not on the interaction. So the Author comes up with some recommendations. The interactions are classified into seven types based on user's intent. I think these categories are not all, most of them are yet to come. The idea of taking user's intent to classify interactions is really a good one.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The paper puts forth some techniques of visualization, such as panning and zooming, overview and detailed interface in a lucid manner. Some of the examples mentioned in the paper are very good and current day applications, such as Google maps. I found the discussion on fish eye effect through the example of Mac OS X icon panel especially, very interesting. Another point worth mentioning is the use of diagrams and the introduction to each type of interface and the subsequent detailed analysis described in the paper.

Reaction : A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The author reviews four techniques of interface design and evaluates the effectiveness of each of them.In the overview+detail, data set is partitioned spatially and there is a brief as well as detailed overview.In zooming, there is temporal separation and the user can choose to magnify/demagnify specific information. In context+focus, the focus is within the context in a single continuous view.In cue-based techniques, only certain items in the dataset which might be relevant to the user are highlighted.

The paper is well written and all the techniques are explained using popular applications like Microsoft Powerpoint, Google Earth that most readers are familiar with. I particularly liked the fisheye views which are explained using Mac OS X Dock icon panel.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The authors have presented four interface approaches in the domain of information visualization: Overview+detail, Zoom&Panning, Focus+context and Cue based systems. It is sometimes confusing as to why the authors consider or report only first 3 categories, but actually they have discussed 4 such approaches.

The authors gave a brief illustration of each of these approaches, their usages, advantages and disadvantages etc., Though the paper is quite informative, it was quite difficult to find out the relevant information, and the organization of the paper is quite confusing.

There are several references and comparisons to the Fisheyes view in the paper. I read some other interesting articles about Fisheyes view from InfoVis wiki, and tried to get a hang of how it actually works. I think, Fisheyes view is very good for usage, and I didn't feel distracted or gone off the target when switching from global context to local context or vice-versa. But, there are several such articles which report of the confusion in using the tool.

Overall, the paper was quite informative, but a little lengthy and confusing. I would make it as a reference, if I intend to research in the direction of user interfaces.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The ideas in this paper are based on the ones described in "The Eyes Have It - A Task by Data Type Taxonomy for Information Visualizations.". These ideas go one after the other, thus resulting in a clear understanding from the user's point of view and are somewhat similar to the waterfall model in software engineering. The seven interaction techniques, which the authors acknowledge cannot be used in all situations, are a starting point for further research in this field. The techniques that a user actually uses will depend on his specific requirement. The authors give justice to the statement "Representation has overshadowed Interaction" through their description of the techniques.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

Information Visualization systems, has two core components: representation and interaction. This paper emphasizes the importance of its relatively neglected component, Interaction.

I agree with the authors' point that, interaction with the information in right ways, will help the users' to gain in-depth understanding, and will help their process of cognition.

Though there are several ways and techniques using which we can interact with the information, the authors made an effort to classify the interaction techniques based on the notion of user intent. The seven categories of classification are: Select, Explore, Reconfigure, Encode, Abstract/Elaborate, Filter and Connect. A brief illustration of each of these categories is provided, which is quite informative.

As authors have rightly pointed, there might be some interaction techniques which might not fall in any of these categories. I think the user-intent based categorization is the right way of categorization, and the authors have laid a foundation stone for research in this direction, and established a meaningful jargon to come up with several holistic frameworks of interaction based on the user-intent.

Reaction: A Task by Data Type Taxonomy for Information Visualizations

As Marchionini rightly said, "the common goals of Information Visualization reach from finding a narrow set of items in a large collection that satisfy a well-understood information need to developing an understanding of unexpected patterns within the collection". The task of exploring the information collections become increasingly challenging as the volume of the information grows. This paper proposes that the useful starting point for designing advanced graphical user interfaces is the Visual Information-Seeking Mantra: overview first, zoom and filter, then details on demand. The paper offers a task by data type taxonomy with seven data types (one-, two-, three dimensional data, temporal and multi-dimensional data, and tree and network data) and seven tasks(overview, zoom, filter, details-on-demand, relate, history, and extracts).

I agree with the authors on several of their arguments. Though, the above types of data serve the research purposes, for successful commercial use the companies have to come up with several novel data structures and several other new tasks apart from the list mentioned above. There were several novel ways of information exploration tools such as fisheye views, but none of them seem to have sustained over time, as they all appear to be fancy in the beginning, but over repeated usage these features tend to not serving the purpose efficiently.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization.

This paper is yet another attempt to express the taxonomies of Information visualization interaction with regard to the other frameworks which are quite subtle and are not much expresssive. According to the authors, Information Visualizations are majorly made of two components - representation and interaction. However , out of these , the representation component has received vast majority of attention in Infovis research.

Corelating Information Visualization with Visual Analytics sounds quite interesting to me. I too think that both of them share common goals and the science of interaction involved, as author mentions, does share similar merits . Though I'm not sure if Visual Analytics should be mostly expressed in terms of data rather than more of visualizations.

The seven categories actually looks similar to the categorizations that were made in the previous papers. I particularly found Reconfiguration interaction very interesting. Author says that by changing the spatial arrangement of representations we can reveal the hidden characteristics of data and relationships between them. As far as what I think, this seems to be quite trivial for reconfigurations like sorting and rearranging colums, however I believe associating reconfigurations with information visualizations ain't an easy task .

For example, let's take a simple bar graph (time vs quantity) representing some statistical data. Now if we reconfigure this, it may first sound good to have linear and sorted representation of bar graphs but on the other hand this may garble the timeline and hence the overall information visualization may not make much sense to analyze.

Overall, I found this paper to be quite helful for anyone to understand and efficiently design interaction for information visualizations.

Reaction: A Review of Overview + Detail, Zooming, and Foxus + Context Interfaces

First of all I found this paper to be quite structured in all respect. Though it is debatable but I can now probably relate that why author has chose single column layout and not double column layout for this.

Once again, this paper mostly discusses the HCI part of the information visualization. Author summarizes the state of research on interfaces that allow users to work at multiple levels of detail and then categorizes a system according to their varying uses of space, time or visual effect.

Particularly with regard to the Fisheye views, the author haven't talked anything about the kind of interaction to achieve them. For example, author has mentioned a little about Fitt's law in HCI. Fitt's law is an important model that determines the time required to rapidly move to a target area. However, most of the Fisheye views that author has discussed would require manual intervention (hovering etc) and if we apply Fitt's law on these views, we can certainly come up with much better and efficient interfaces. Whether it is a trade off or something useful, we can debate on that.

I really liked the way author has explained the four categorization of system with the help of real world examples of Google maps and other commonly used desktop applications. However I don't think that the kind of interfaces i.e. Overview+Detail , Zooming, Focus+Context would be enough to capture most of the modern interfaces that are emerging these days, for example interfaces for multitouch tablets and phones.

Reaction : A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

This paper extensively discusses about different types of visualizing schemes. They are,
Overview+detail interface: Here the data that designer wants to represent is done in a very detailed manner.
Zooming: the advantages of zoom while cognition of visualization. Though easy to implement, it requires lot of cognition from the user, as once zoomed in, the user has to keep track of where he is, in terms of navigation.
Focus + context: As the name suggests this is more about focusing on specific information based on the context.
Cue: This scheme concentrates on minimizing the size by modifying the data.

The description about panning and scrolling and the advantages of above methods over it is very well explained. The above mentioned schemes are sometimes not of any help when used alone. When used together they serve the purpose of better visualization. I found that there is a lot of overlap between the visualization schemes that the author categorized. The examples that the author provides are very fresh and closely relate with the current day application, which made my understanding of the topic much clear.The advantages and disadvantages of the schemes was well put.

Sunday, September 18, 2011

Reaction : Toward a Deeper Understanding of the Role of Interaction in Information Visualization

A very clear distinction between interaction and representation is presented in the paper and a summary of the amount of research that has gone into both the areas is described. Interaction and representation go hand in hand but neither of them can be compromised for the sake of the other. The paper also gives us the chance to ponder about the fact that representation of infoVIZ has always overshadowed the interaction aspect. The paper makes an effort to bridge in this gap by describing the seven interaction techniques.

The interaction techniques which are described are detailed in nature but I felt that each interaction technique could have included broader range of examples. The interaction technique which we actually apply depends on the users requirement and thus cannot restrict ourselves to the listed techniques, that is the technique to be applied is rather "tightly bound" to the users requirement. The authors acknowledge this point and mention that there is a large scope for further research this field.

Overall, even though the paper cannot formally state that these seven techniques can be applied in all situations, it puts forth the extensive research conducted in this field and provides the reader background knowledge which will help pave the way for extending the work on interaction techniques.

Reaction : Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The paper clearly gives the relation between representation and interaction. The paper is more biased towards dynamic visualization than static, which I feel should also have been included in his definition on visualization. After reading the paper "The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations", I can closely relate the taxonomies and their practical implementation. The author supports his reason to stress on interaction more than representation by stating that enough research had already been done on representation that interaction, which I strongly agree.

The categorization of interaction was brilliantly done, using a rationale scale. The examples that author provides to each of the categories helped me in understanding the concept really well. The start of the paper was done elaborately by giving the synopsis of the research done till now, by summarizing some popular papers in the field. The screenshot's are blurred. I think its because of the size constraint. This paper is a pathfinder to the research on interaction.

Reaction : The Eyes Have It - A Task by Data Type Taxonomy for Information Visualizations

The paper discusses several tasks and data types that aid readers in visualizing a visualization. The examples provided for each data type gives the reader a good idea about what the author wants to convey in his theory. The traditional relational databases can be replaced with good visualizations is one of the point that the user makes, but I feel that when it comes to representing huge data, nothing can beat relational databases. The author's summary of the basic visualization mantra " overview first, zoom and filter, then details-on demand" reiterates the point of KISS(Keep It Simple Stupid). The detailed analysis of the tasks is very helpful in understanding the concept, and the advanced filtering options is an interesting topic to do research on. I completely agree with the author that any data can be visually represented by the seven data types. the relation of the tasks with the data types was explained in a very cognitive manner and helped me in understanding the difference between the data types clearly.

Overall the paper looked a bit outdated as the examples provided in it are quite far from the current day implementations of the same tasks and data types. But, the concepts are still the foundations of the present day theories that details about interaction with the visualization. This paper mainly helps a designer while developing his visualization as most of it deals with the ways that help a reader in visualizing the visualizations.

Reaction: The eyes have it: a task by data type taxonomy for information visualizations

I feel this is yet another very good paper which talks about the importance of information visualization and how it replaces the old way of representing data (database management). The author also emphasizes of "bandwidth of information visualization" to give the user a reason to learn more about visualizations. The author asks the UI designers to keep in mind the power of the human eye and its ability to distinguish various colors and shapes.

Mainly the paper gives us a "mantra" which tells us things to keep in mind depending on their relative importance. I found "task taxonomy" to be very interesting because it tells us the various dimensions which we need to keep in mind when visualizing data. Clearly the author makes an attempt to speak out to a broader range of audience including those from relatively non-technical backgrounds by describing different types of dimensions and giving examples to show how they can be visualized, thus creating an "abstraction of reality".

It can be clearly seen that the principles mentioned in the "mantra" are applicable even till date and some of the best user interfaces have ensured that sufficient effort is spent on describing each principle in a detailed level. The author then goes on to describe the "Task by data type taxonomy" with sufficient examples.

Summarizing my thoughts, I feel the author has done a very good job in describing the principles in a detailed fashion but somehow the information presented, especially the sub-headings could have been organized better to create a more efficient information flow process. The paper has an abrupt ending without a smooth transition from "Advanced filtering" to "Summary". This is definitely a good read and I will follow up with some of the references mentioned in the paper.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The author of this paper has emphasized on a very good point i.e the interaction aspect of information visualization has been overshadowed by the representational aspect.Moreover, representation and interaction both go hand in hand for a successful information visualization. One is incomplete without the other. There still is tremendous amount of scope for research in interaction aspect of information visualization. This paper is an attempt to encourage more research this field.The various interaction techniques have been effectively described.

After tremendous amount of research the authors have classified interaction techniques in seven categories. The content is extremely informative and each of these categories have been explained thoroughly with good illustrations. However, the interaction techniques might straddle between two categories. It is not always possible to classify a particular technique under one category. Interaction is an important aspect of visualization. In order to achieve success, the user intent should be given more weight. In other words, the focus should be on the user intent rather than how a particular technique provided by Info Viz. works.

Reaction:The eyes have it: a task by data type taxonomy for information visualizations

In this paper, the examples are really good. It clearly distinguishes what data-types should be used. The Tree and the Network data type would have been really confusing if the examples were not provided. Advanced filtering with dynamic queries, boolean expressions, venn diagrams are few unfamiliar topics.


Reaction:The eyes have it: a task by data type taxonomy for information visualizations

The paper starts with importance of Visualization, and how it pushed aside traditional methods like Relational Databases. It stresses on one basic technique which it calls as a mantra " Overview first, zoom and filter, then  details-on-demand" and tries to explain, justify it. The author also discusses about different data types classified as ( ( 1 - ,   2-, 3-dimensional data, temporal and multi-dimensional  data, and tree and
network data). The paper discusses each type and talks about the part that user wants to see and different Visualization methods adopted to represent the data. The paper then goes on about each task in the mantra and concludes that the current novel Visualization techniques should adopt this mantra and get better. 

Reaction: The eyes have it: A Task by Data Type Taxonomy for Information Visualizations

This paper tells about the various forms of data types. It talks about the visual information seeking mantra and hows the perception of visualization is complete due to it. There is well defined characterization of the practices in information visualization.

The author draws attention to history and extraction. It is interesting to see how seeing historical information on being analysed leads to better information visualization.

The authors view of taxonomy and tasks is very efficient in analyzing complex graphical visualizations.

Reaction: The Eyes Have It - A Task by Data Type Taxonomy for Information Visualizations

This article does a great job of organizing the different types of interaction and data types one might encounter when interacting with a visualization. I believe that the Multi-dimensional and Network data types provide designers with higher opportunities to allow interactive manipulation of the data represented. The authors of the article add that the increased interaction requires users to practice before being able to fully comprehend.

The authors also point out that the Tree data type has been re-represented into indented structures. This movement makes sense in terms of representing the data in a compact form, but if it takes users 10-20 minutes to understand the data, perhaps this type of tree is not worth the trouble.

I found the seven tasks described in this article to be very helpful. In "Toward a Deeper Understanding of the Role of Interaction in Information Visualization" the authors list several specific interaction techniques in a somewhat unorganized manner. The authors of this article choose to point out the most significant tasks the user may want to take. The frequency of these tasks provide enough motivation to include their functionality into any InfoViz project.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

True. Interactions are very important for InfoViz and makes the user involve more into the visualization. Paper clearly defines 7 types of interactions in which I feel that reconfigure is not so essential as an interaction. Because Encode interaction helps the user to change the way how the data is represented. I am not sure how many users will be actually performing a reconfigure.


Reaction:The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations

The paper gives a starting point that will aid the designers of advanced graphical user interfaces. According to the author, the data is classified into a taxonomy of seven different types and the tasks that needs to be performed is again divided into seven types.The author asserts that a better design can be developed to visualize data by taking advantage of the under utilized human perceptual abilities. The article stresses on the visual seeking mantra of overview first, zoom. filter and details on demand.

 Overall the article has been systematically organized by explaining the various data types, its advantages and disadvantages. This is ensued by briefly explaining each component of the visual seeking mantra. It is a very good read for designers of an Information Visualization tool. The bifocal display representation and the filter flow model for dynamic querying was a very good piece of information. According to me, The Visual seeking mantra, if applied in a prudent way will help in increasing the insight of the end user in optimum number of steps.

Are the steps 'details on demand' and 'extract' related? I was not able to infer a very clear distinction between them.

Reaction: Toward a Deeper Understanding of the Role of Interaction in Information Visualization

The different views presented in this article of what interaction encompasses give insight into the many ways users can interface with a visualization. I agree with the authors' definition of interaction as the ability to manipulate the interface to change the data represented by the visualization. This definition excludes any sort of static info graphics from the field, but I feel that they should still be included. Info graphics still provide viewers with data that can be interpreted without interaction.

The authors describe the separation between static and interactive visualizations as adding two different roles: user and viewer. The taxonomies of interaction techniques presented in the article show that role of the user is a much more interesting role to take. The descriptors for each of the different techniques show that there can be several unique methods of interaction that may only slightly differ from others. For example, I feel that categorizing and filtering could be describing similar interactions with the exception of categorization presenting several different sets of related data on the same view and filtering only displaying one.

Overall, the authors of the article give a good representation of the most common interaction techniques. Most importantly, they help define the goals users have in mind when performing these techniques. These goals help designers choose the proper tools to provide users data manipulation in an info visualization.

Reaction: A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The author focuses on 4 approaches for working in interface schemes namely overview + detail : which uses spatial separation between focused and contextual views , Zooming : which uses temporal separation, focus + context : which minimizes the seam between views by displaying focus within the context and cue-based : which selectively highlight or suppress items within the information space.

Discussion about resolution of eyes and spatial partition, discussion about fish eye effect with example of Mac OS X were good.Examples of google map for overview + detail and  Microsoft power point were nice examples.  The perspective wall which was not covered in the paper The Eyes Have it ,  where we read and discussed about 7 basic taxonomies were discussed in this paper which I found very useful. They have tried to generalize the fish eye view using the formula degree of interest. fish eye tables, fish eye documents etc were good to categorizations too.

Even though the author says target acquisition in Mac OS X for dock is difficult with the fish eye view i find it more appealing to the eyes because of the animation which is different from the usual dock stations. May be in other cases what the author says is agreeable. Even though author has categorized four approaches , I felt all of them are mutually exclusive and can not be categorized as author has mentioned. Overall the paper gave lot of information and new learnings and was a wonderful read.

Event: NCDevCon 2011

NC State was host to a very popular developer's conference this weekend. Organized by the Triangle Area ColdFusion User's Group, it was a two day event on the Centennial Campus which I enjoyed attending. Big names from Adobe were present. Variety of lectures on mobile development, web development and cold fusion were arranged. Some of the sessions that I attended and found very interesting and relevant were 'HTML5 deep dive', 'Introduction to jQuery Mobile', 'QR Code Crazy' and 'Geolocation 101'.

The official site and video recording of lectures are available on this site:



The title of the paper itself got me interested. 'The Eyes Have It', though written way back in 1996, seems so relevant even in today's context. Simple concepts, explained lucidly seem almost intuitive. In fact, it was exactly how I approached the course project when we started low-level prototyping. I think this paper is a 'classic' for information visualization. Even then, Shneiderman anticipated how visualizations would play a major role with advent of high definition color displays and ever improving technology. Another related work that I found useful was a paper on 'A Task Oriented View if Information Visualization', written by Hibino at Bell Labs

(link to paper:

I think the concepts put forth in this paper can also be adapted to current technologies. I particularly would like to put this in context of mobile displays or displays for tablets. There are so many points of interaction with the device for users. The number of modalities increases with the 'touch' concept. Even then, the 'mantra' as Shneiderman says, "Overview first, zoom and filter, then details-on-demand" remain the same! Off topic, similar to this paper, there is also a paper titled "The Ears Have It - A task by information structure taxonomy for voice access to web pages"

A Review of Overview+Detail, Zooming, and Focus+Context Interfaces

The paper has shifted focus from plain study of interfaces to zero down on the effectiveness of these interfaces. I like the way the paper was organized with enough introductions to each type of interface and then a detailed analysis based on the survey of the others in the subsequent sections. The use of diagrams in the papers was especially helpful and I liked that. I also agree with the authors that motives can be categorized as target acquisition and a general comprehension of the information space depicted. Both are effectively expressed through different types of visualization. I also particularly liked the concept of ‘value bars’ mentioned in the paper.

Toward a Deeper Understanding of the Role of Interaction in Information Visualization

Interaction no doubt, has always been a part of information visualization. What the paper aims to do is extensively study interaction/interface techniques used in visual analytics, identify areas which are not effectively represented and suggest interaction techniques, evolve existing ones to “fill those gaps” as the authors put it. The inclusion of user intent quite natural and the seven categories seem to cover the intent portion very well.