A Comparative Analysis of SequoiaView and Grokker: Information Visualization Tools, Assignments of Computer Science

Download A Comparative Analysis of SequoiaView and Grokker: Information Visualization Tools and more Assignments Computer Science in PDF only on Docsity! CS 7450 Assignment 06 Tarang Taunk gtg180j (Submitted on: March 15, 2005) Introduction In this report I will present a critique of two information visualization tools – SequoiaView and Grokker. Although the visualizations offered by these tools appear very disparate, both systems are based on the ‘space-filling’ representation algorithm called the Treemap. I have evaluated both tools for the quality of information conveyed by their visualizations and for the usability of their user interfaces. Critique of SequoiaView SequoiaView is a visualization tool used for searching and browsing through the files and directories in a computer on which the tool is loaded. To understand how the tool functions and to explore all the features available, I set some goals for myself that I would try to accomplish using this software. I made sure that the tasks I performed are a good representative of a user’s goals browsing through files and directories in a computer. Following is a list of the tasks I performed and a brief description of my interaction with SequoiaView: Interaction with SequoiaView: Determine the largest directory (in terms of hard disk space used) on my computer: This task was very easy to perform and I determined the largest directory without doing anything at all. SequoiaView automatically loaded C drive in its window and the largest directory on my computer was evidently “Portfolio” in “My Documents” in “C:” As seen in Figure 1, the largest box of files in the visualization was the one somewhat in the center with multiple pink boxes within it. Hovering the mouse on top of these gave the entire path of the files in this folder. Fig. 1 Determine the most common file type in My Documents: My guess was that being a student of architecture and having taken various classes dealing with graphics, I would have more image and video files than text files saved in My Documents. I tried to verify this hypothesis by changing the default color scheme provided by SequoiaView. I assigned only three colors to the new color scheme – red for all types of graphics files (including images and movies), blue for all types of text files, and all other files defaulted to gray (see Figure 2). Figure 3 shows the result. At first glance, it seemed like there were many more “red files” than “blue files” or “gray files”, thus confirming my hypothesis. However, on closer inspection, I could tell that the screen was covered more with red area than either of the other two colors. But I could not accurately say if the files in red were more in number than the ones in blue or gray. Most of the boxes representing blue and gray files were much smaller in area compared to the red files. This was quite understandable considering the fact that image and video files tend to occupy a lot more memory space than text or exe files. The area distribution among the red, blue and gray boxes had initially biased me. A closer look at Figure 3 will reveal some nested folders that contain small, but a large number of gray and blue files which might add up to be more than the sum of all red files. SequoiaView does not provide counts of various types of files and hence, I could not confidently tell which the most common file type in My Documents was. Determine the most common file type in My Documents: The second task was to determine the most common file type in “My Documents”. Although Grokker allows choosing between various “color themes”, it does not allow one to color code say the image and text files separately. Hence, this task was not as easy to accomplish as in SequoiaView. However, Grokker did allow me to filter files of a given extension (all jpeg files, for example) and hence enabled me to view populations of one file type at a time. Find the location of a file called “Tarang.jpg”: Unlike in SequoiaView, the task of finding the file “tarang.jpg” was extremely simple and trivial in Grokker due to its ability to search for a particular file. Figure 7 shows the location of the file. I found that there were actually two files on my computer with the name “Tarang.jpg”. Search for prices of Cannon Digital Cameras sold by Amazon: The search for cannon digital cameras revealed various categories of search links which contained pages with those key words. It was not clear to me how the different categories were formed by Grokker, however, navigating through them was fairly easy. Zooming into a category named “Digital Cameras” I could see thumbnails of various cannon cameras and clicking on one of the thumbnails opened the Amazon page describing the camera and its price in the panel on the right. I could further filter the search results based on maximum price and customer rating. Figure 8 shows a selected camera and its Amazon page open. Fig. 8 Fig. 9 Learn more about MEMS using Google: Figure 9 displays a search on the word “MEMS” using the Google engine. As in the previous task, Grokker returned various categories of links to choose from. Since the goal of this task was to simply browse through the pages to learn more about the MEMS technology, I filtered the results based on “Rank” to see only the top ranking pages on the subject. When I hovered my mouse pointer over the links, I could see a synopsis of the data in that page. I used this information to browse through various links and clicked on the links that I found interesting. Grokker – my findings: 1. My first impression of Grokker’s visualization was very good – the images of squares and circles nested within others conveyed the hierarchical structure of files and folders very efficiently. However, the extension of this principle to web-based searches was not intuitive and I found myself constantly referring to the Help files while analyzing the results of testing the software for web searches. 2. Grokker is a much more powerful tool than SequoiaView and hence needs a user interface that has been carefully designed to enable its user to utilize the software to its maximum potential. However, I found the interface complex and unintuitive. 3. The software incorporates a powerful search engine that makes searching for files very easy. If one has a rough idea about the layout and structure of the tree, Grokker can also serve as a fair browsing tool. Although, it lacks in giving an accurate “overview” of the files and folders based on size like SequoiaView does, it does provide a fairly accurate “overview” of the nested structure. 4. The size of the circle representing a folder in Grokker seems to be proportional to both, the memory it occupies and the number of folders and files contained in it, simultaneously. This makes it impossible to obtain comparative results of a folder’s memory usage or the count of the files within it. 5. Grokker is very successful as a search and browse tool to look for items on the internet. Grokker is great for searching items that have specific image files associated to them because the tool displays the thumbnail within the visualization, making the search easy. When browsing sites such as Amazon and Google with a regular internet browser, it is very easy to click on a few links and easily wander away from pages of interest. Grokker has an in-built solution for allowing the user to return to the “overview” level, thus refraining the user from getting “lost”. 6. The methods used by the tool for creating the categories of the web pages are not transparent and hence might confuse the user initially. I liked that a single page was cross-referenced in different categories if it fitted under more than one category. However, again, it was not clear how the system made decisions regarding the location of the pages under different categories. 7. Although the graphically rich interface is aesthetically pleasing, it places an extremely heavy burden on system resources, leading to a slow performance that is sometimes very inconvenient.