Comparison of Information Visualization Softwares: SequoiaView and Grokker | Assignments Computer Science

Homework 4 Punit Gupta

CS7450 Information Visualization [email protected]

Introduction

This assignment is aimed at comparative study of two visualization softwares namely,

SequoiaView and Grokker. Both of these softwares take a different approach towards

visualizing the same hierarchical information set. This document is organized in the

following way, first I will give a brief introduction of the approaches employed by these

applications, which will be followed by a comparative study of the strengths and

weaknesses of these softwares which includes a short discussion on the UI of the

softwares.

Grokker

Grokker uses a circular (user has the option to change that to rectangular representation)

representation for visualizing the hierarchical dataset. It employs zooming to navigate

through a large amount of information set. As shown in Figure 1 Grokker lays out the

directory hierarchy structure such that user can see the directory names and it’s contents

in a clear fashion. Individual items are arranged in the alphabetical order. In addition to

that option for zooming in/out makes it appropriate for navigation purposes. The tool can

also be used to visualize the search results either from disk or from internet. Grokker is

also known as visual search engine.

SequoiaView

SequoiaView is based on the Treemap visualization technique. It is a space filling

visualization, where space occupied by an entity is directly proportional to its size. This

makes SequoiaView more suitable for providing the user with an overview of the dataset

and identifying clusters of information which is not obvious in the traditional tree like

visualization of folders in windows. Also the systems visualizes the information set in a

very abstract form which does not carry in itself lot of information required for

navigation in the information space. Hence unlike Grokker, this is not very helpful for

navigational purposes. This could prove to a powerful tool for the users who are

concerned with the space management in the disks. For example when I visualized the C

drive of my computer with SequoiaView, it resulted with the visualization shown as in

figure 2. Just by looking at the visualization one can make out the fact that there is a quite

a big file being present in the disk which is shown by the gray rectangle. Looking at the

details at the bottom of the screen tells that it occupies 13.2% of the whole disk. Being

led by the curiosity I moved the mouse over the rectangle and it displays the name of the

file as tool tip which is pagefile.sys.

Comparative study

1. Grokker’s zooming feature seems to be very powerful when it comes to navigating

through a complex hierarchical structure. User has the option of drilling down into the

Partial preview of the text

Download Comparison of Information Visualization Softwares: SequoiaView and Grokker and more Assignments Computer Science in PDF only on Docsity!

Homework 4 Punit Gupta CS7450 Information Visualization [email protected]

Introduction

This assignment is aimed at comparative study of two visualization softwares namely, SequoiaView and Grokker. Both of these softwares take a different approach towards visualizing the same hierarchical information set. This document is organized in the following way, first I will give a brief introduction of the approaches employed by these applications, which will be followed by a comparative study of the strengths and weaknesses of these softwares which includes a short discussion on the UI of the softwares.

Grokker Grokker uses a circular (user has the option to change that to rectangular representation) representation for visualizing the hierarchical dataset. It employs zooming to navigate through a large amount of information set. As shown in Figure 1 Grokker lays out the directory hierarchy structure such that user can see the directory names and it’s contents in a clear fashion. Individual items are arranged in the alphabetical order. In addition to that option for zooming in/out makes it appropriate for navigation purposes. The tool can also be used to visualize the search results either from disk or from internet. Grokker is also known as visual search engine.

SequoiaView SequoiaView is based on the Treemap visualization technique. It is a space filling visualization, where space occupied by an entity is directly proportional to its size. This makes SequoiaView more suitable for providing the user with an overview of the dataset and identifying clusters of information which is not obvious in the traditional tree like visualization of folders in windows. Also the systems visualizes the information set in a very abstract form which does not carry in itself lot of information required for navigation in the information space. Hence unlike Grokker, this is not very helpful for navigational purposes. This could prove to a powerful tool for the users who are concerned with the space management in the disks. For example when I visualized the C drive of my computer with SequoiaView, it resulted with the visualization shown as in figure 2. Just by looking at the visualization one can make out the fact that there is a quite a big file being present in the disk which is shown by the gray rectangle. Looking at the details at the bottom of the screen tells that it occupies 13.2% of the whole disk. Being led by the curiosity I moved the mouse over the rectangle and it displays the name of the file as tool tip which is pagefile.sys.

Comparative study

Grokker’s zooming feature seems to be very powerful when it comes to navigating through a complex hierarchical structure. User has the option of drilling down into the

information buried beneath multiple levels thus exploring the detailed view of a portion of the information hierarchy and zoom out in order to see the overall picture of the dataset. Though SequoiaView also provides options to zoom up and down a hierarchal data set, it is not as intuitive as Grokker. Grokker presents the zoomed rendering of the dataset within the context, while SequoiaView does not retain the context when zooming into a particular directory.

In SequoiaView as the size of the rectangle decreases it becomes harder to distinguish one directory/file from another. Sometimes it might fail the sole purpose of the visualization. In Grokker this situation is even worse because Grokker is supposed to be more suitable for navigational purposes. In case of really small cells, user may not be able to click on it and see the details. SequoiaView provides the physical zoom (enlarging the canvas rather than displaying its children nodes) to help alleviate the problem of information clutter.
Though Grokker provides ZUI like navigation, it does not have support for panning or scrolling. That is necessary for the users to navigate through the information space. For example if the user clicks on a particular directory say A, its content is displayed. User clicks on a directory (X) from the content view, it brings the node into focus and the system displays all its children files and directories. Now if the user wants to see other files/directories which are sibling to X, he would have to zoom out a level and then search for that item. However a natural way to explore the dataset is to zoom and pan interchangeably as needed.
In SequoiaView if you move the mouse over an item, it shows the fraction of space occupied by that item in the whole disk as well as in the directory. Grokker does not have this feature. Small but could be helpful in some instances where user is interested in learning the pattern shown by the files/folders in terms of occupying the space in the disk. For example in Grokker there is no way to answer this question, “which file extension occupies the maximum space in the disk?”. In Sequoia it is almost apparent to answer such questions. It turns out that in my case around 1/3 of the disk is occupied by mp3 files. This result has been obtained by a close observation of C drive in SequoiaView. Looking at Figure 2, one can see a large chunk of green cells (representation for mp3 files) occupying almost 1/3 of the disk.
SequoiaView seems to be a powerful tool for information clustering and identifying abnormalities in the data set. For example in Figure 2, green color shows the mp files and in between there are few rectangles with gray colors which represent files which windows does not know how to open. For example I had few of my favorite songs in .mpga format, and Windows does not know how to play this format. This led me to research about the codec which might be needed to play this song, and apparently it turned out that all I needed to do was to rename the .mpga files to .mp3. I also came across a very interesting serendipitous incident while exploring my dataset with SequoiaView. While browsing through the song folder (mostly green files) I saw a gray colored rectangle, which happened to be the song I was looking for quite some time. It’s “please forgive me” from Brian Adams. I remembered

Figure 2, where green stands for MP3 files. Grokker lacks this feature of distinguishing a file format from another.

Grokker is very flexible in terms of displaying the filtered items. It provides three ways to visualize items after a filter has been applied,
- Hiding the filtered items
- Making the filtered information invisible
- Graying out the filtered items This feature can be very useful in the information spaces which are very large, and would require user to play with the filtering feature. On the contrary SequoiaView does not have any flexibility on this front.
Grokker implements the notion of direct manipulation for zoom in/out feature, meaning user clicks directly on the directory he wants to focus on (object of interest). On the contrary in SequoiaView there is a spatial distance between zoom button (to go up in directory) and the object of interest (directory).
In Grokker sliders for color coding the map based on the filter applied is very powerful. However sometimes it becomes hard to figure out the values of the variable represented by a particular shade of color. A possible solution to this problem could be to label the color slider for each filter, which would give the user an idea about the value of the variable for that color, it would help even more if the system shows the value of the variable as tool tip, as the user moves the mouse over a particular color value.
In Grokker I am still wondering about the color coding in the original visualization. Before applying any filtering it seems to have color coding for the items. I was prompt to think that the files might be color coded based on the size, but I could not figure out that thing.
Grokker is misleading in terms of portraying the size of the folders through its visualization. The size of the balloons on the screen is not proportional to the size of the file/folder. This might prove to be a false conclusion.
In Grokker it is not very intuitive to figure out the function of the top-center slider which controls the value for details. If set to show less details, it hides away few items according to some criterion which I do not seem to understand. Further the icons used at the both ends of the slider are normally used for physical zoom feature. So I was expecting the same behavior from it, while it seems to implement semantic zooming.

Conclusion:

Overall I really liked the idea of visually navigating through the hierarchical information set, the paradigm promoted by Grokker. However it does not seem to provide insights in terms of the size of the items. SequoiaView does a good job of providing the overview of

disk usage; however it is not very useful in terms of navigating through the information set.

Figure 1: Rendering showing the visualization in Grokker

Figure 2: Treemap based visualization in SequoiaView

Comparison of Information Visualization Softwares: SequoiaView and Grokker, Assignments of Computer Science

Related documents

Partial preview of the text

Download Comparison of Information Visualization Softwares: SequoiaView and Grokker and more Assignments Computer Science in PDF only on Docsity!

Introduction