Download Exploring Information Hierarchies: A Comparison of SequoiaView and Grokker and more Assignments Computer Science in PDF only on Docsity!
Homework 4: Examining systems for exploring hierarchies
Paula Edwards 2/23/
For this assignment, I examined two tools, SequoiaView and Grokker, which are used for visualizing and exploring information hierarchies. SequoiaView uses the treemap visualization technique to display information about files and folders on a disk. Grokker uses circular visualization with zooming to display information about files and folders on a disk or information from the web. A review of each tool is provided below followed by a comparison of the two tools.
SequoiaView SequoiaView provides a good overview of the distribution of files among various folders on the disk or parent folder of interest. It was especially well suited to understanding how space is distributed on the disk. Using the file-type color scheme, the user can easily see which file types occur most frequently, which files/file types take up the most space, and which file types have most variability in amount of space. This information could be useful for disk maintenance activities such as deleting or archiving files to make room for new material. It might also be useful for system administrators for understanding which software a user or user group uses (based on the filetypes), but looking at the files for a large number of users would be time consuming and possibly tedious.
The ability to color files based on their type really added to the usefulness of the visualization. Because the color scheme is so crucial to the visualization, it is good that the user can customize the color scheme based on the type of files they frequently use. For example, I use SPSS, a statistical package that creates has large files data and output files. These files have different file types (.sav, & .spo) from those provided in the default color scheme. By modifying the color scheme, I could highlight these files. As a side note, I didn’t really like the initial color scheme. It seemed a bit too random, although, using various shades of blue for all MS Office files was nice since these are logically similar types of files. In using my modified color scheme, I was able to see that most of my files (approximately 70-80%) are MS Office, SPSS, or pdfs (see Figure 1). One thing that I disliked about the coloring was that when you close the modify color window, the application does not prompt you to save your color changes. Therefore it is easy to forget to save them and you have to reapply the color changes the next time you open SequoiaView.
Figure 1: SequoiaView with modified color scheme (Office files in blue & gray-blue, pdf green, SPSS yellow)
In general I found that is was not a good navigation aid. For example, it was difficult to find a specific file or examining the content of a specific file. The mouseover and status bar give the file name, which is nice, but looking for a specific file required mousing over a number of squares at random until you found the right folder, then file. I tried several different methods of ordering the blocks (e.g. from order by size to order by name), but that only slightly improved the ease of this task. Also, the visualization did not help the user understand the folder structure to support more effective navigation. By default, the system provides a yellow highlight of main folder the file that is ‘current’ belongs to, but you cannot easily tell what the highlighted folder is. Specifically, the mouse over shows the path, but you don’t know which parent folder in the path is being highlighted. For example, in Lab\Miami 2 Study\Data\data.xls, is the highlighted yellow folder Lab or Lab\Miami 2 Study? Changing the options to highlight ‘File and entire path’ helped clarify the navigation path for the selected file (see Figure 2.), but big picture view of navigation was still missing.
Figure 2: SequoiaView with highlight file and entire path enabled
Filtering was useful for answering more specific questions regarding file size, usage, or type. Filtering on size and date were especially useful, and I liked that you could change the direction for the threshold (file accessed before X date vs. after X date). This type of filtering enabled me to gain the insight that there are a lot of files on my hard drive that haven’t been accessed in over a year. Unfortunately, filtering by file type did not work for me. I suspect that this may be due to the fact that the application changes all file types entered in the filter box to all caps and all of my file names use lower case letters.
In addition to filtering, SequoiaView provided zooming. However, I’m not sure how useful it is since when you zoom you loose the overall context. The only advantage I see to using zoom is browsing when you have lots of little files.
Grokker Grokker was quite useful for understanding and navigating a hierarchical information set, although less useful for examining space allocation. Locating and navigating to a specific file was quite easy. The nested circles provide an overall context that is very useful in understanding the navigation structure. The size of the circles provides a visual cue as to the number of files/folders within a directory and the shapes differentiate folders from files. Browsing was especially easy in the split view that showed the Grokker circles on the left and a browsing space on the right. In this view clicking on a folder in the Grokker window displayed the Explorer view of that folder. For pictures, a thumbnail was also provided in the Grokker window. I found this combined view easier to use for browsing
Figure 4: Grokker filtering view
After using Grokker to browse my hard drive, I also tried to grok a website, www.isye.gatech.edu. I found this quite confusing. First, grokking was slow, so I was a little worried that I had done something wrong, like trying to grok something that is too big. Then, when the results were displayed, I was not sure what I was looking at. I expected to get something like a site map of the ISyE website. Instead, I got what seemed to be hierarchy of websites that link to www.isye.gatech.edu. (See Figure 5.) In looking at the filter options, there was a ‘Source’ option. In looking at the sources, I began to believe that the systems takes the results from a search engine or directory site, like Yahoo, and pumps them into Grokker. I am still a little unsure where the navigation hierarchy comes from, unless one or more of the sources have a classification scheme of some sort. That being said, my confusion probably resulted from my misconception of what the results would be. I thought that entering a web a would let me browse the files on that site the way I browsed files on my hard drive. Actually, web grokking searches the web and organized the results based on a hierarchy, though I’m not quite certain where the hierarchy comes from. I could see how this would be a useful way to browse web search results, but it will only be effective if the classification scheme used to create the h of search results is effective.
ddress
ierarchy
F igure 5: Grokker view of www.isye.gatech.edu
Comparison of SequoiaView and Grokker SequoiaView and Grokker take two very different approaches to visualizing hierarchical data, therefore, they were very different in the types of tasks they support. SequoiaView is better suited for tasks the examine file space and file type distributions in a directory. Grokker, on the other hand, is better suited to browsing and navigating a hierarchical data structure. For example, Grokker was better than SequoiaView for searching for locating a file. On the other hand, Grokker does not give you the overall picture of space allocation and distribution of file types that you get with SequoiaView.
Both systems provided some form of filter and zoom capabilities, although I thought the filter and zoom in Grokker was more effective, given its usefulness in browsing and navigating. However, SequoiaView’s filtering was more robust in that you could set the direction of the filter (above vs below) in addition to the threshold. If Grokker provided similar filtering capabilities as SequoiaView, it could be quite useful for file management, especially in knowledge management applications. SequoiaView, while well suited to its intended task, I found less useful. I find that in my day to day activities, I have little need to examine space utilization on my hard drive. Of course, I know several system administrators that would probably argue with me on that point.
