Exercises

Chapter 1: Introduction

  1. Choose a topic from computer graphics or visualization and research its origins. Feel free to skim ahead in this book to find a topic, such as volume rendering or parallel coordinates. Send your contributions to the authors via the book’s web site. If we can verify it, your findings may be placed in our online history page.
  2. Describe the linkages between the stages of the graphics pipeline and those of the visualization pipeline. Are there any stages in one pipeline that do not have a clear linkage in the other pipeline?
  3. Describe the linkages between the stages of the visualization pipeline and those of the knowledge discovery pipeline. Are there any stages in one pipeline that do not have a clear linkage in the other pipeline?
  4. Give an example of each of the three categories of visualization: presentation, confirmation, and exploration.
  5. Familiarize yourself with scatterplots: write up a summary of what they are, how they are created, and how they are used. There are hundreds of different variations on scatterplots, so select one as an example in your summary.
  6. Select a member of the MIT aesthetics and computation research group (http://acg.media.mit.edu/). Briefly discuss that person’s work and provide a review of the potential for that technique to help in information visualization (amount of information communicated vs. amount of aesthetics).

Chapter 2: Data Foundations

  1. Give examples, other than the ones listed in this chapter, of data sets with the following characteristics:
    1. with and without an ordering relationship,
    2. with and without a distance metric,
    3. with and without an absolute zero.
  2. Describe the difference between a data attribute and a value. Use examples to clarify your response.
  3. There are numerous strategies for dealing with missing data in a data set. These include deleting the row containing the missing value, replacing the missing value with a special number, such as −999, replacing the value with the average value for that data dimension, and replacing the value with the corresponding entry from the nearest neighbor (using some distance calculation). Comment on the strengths and weaknesses of each of these strategies: what is gained or lost by following one approach over the others?
  4. Perform a web search looking for repositories of publicly available data. Retrieve two or three, and analyze them in terms of their structure and meaning. Does the data have spatial or temporal attributes? Is it nominal or ordinal (or both)? Does it come in a standard or custom file format?
  5. Repeat the above process, using a newspaper as your source. What sorts of data can you extract from the newspaper? What are the data types? What data sets could you derive by processing the information in the newspaper? Try to design at least one data set for each section of the newspaper.
  6. List at least ten sources of data from your normal daily activities (you’ll be surprised – data is all around us!). For example, nutrition labels from the food we consume have a wealth of information, some of which you probably don’t want to know. Start gathering one or two types of data to be used for future projects in this course.

Chapter 3: Human Perception and Information Processing

  1. List the features you believe you use in recognizing a friend by sight and/or by sound. How might you use related features to communicate a data set?
  2. Design an experiment that would integrate an eye tracking study with a target discovery test.
  3. Design an experiment to identify which is better for visualizing a linear pattern in a large data set: a simple point plot, or a point plot where the points are circular, rectangular, colored, or vibrating. Guess at the outcome.
  4. Since about 8% of males are color deficient [235] (with less than 1 % for females) mostly in the red and green ranges, how would you deal with color in the display of a scatterplot?

Chapter 4: Visualization Foundations

  1. Show that Mexp and Meff, as defined earlier, are distance metrics.
  2. Identify some of the tools, systems or packages listed in Table 4.1 that are either outdated or no longer available (lots of visualization companies have come and gone!).
  3. Identify and describe some currently available visualization tools, systems, or packages that could be added to Table 4.1.
  4. Compare and contrast two or more of the taxonomies or classification schemes described in this chapter. Choose the ones you feel have the most overlap.

Chapter 5: Visualization Techniques for Spatial Data

  1. These days, many people carry small, portable display devices with them, such as mobile phones and PDAs. Discuss the ramifications of migrating to a small display for the visualization techniques discussed in this chapter. Which mappings maintain most of their benefits when scaling occurs? What strategies might you pursue to allow viewers access to the same or similar resolution of information?
  2. Researchers in the visualization field have spent considerable time trying to differentiate classes of techniques, such as scientific versus information visualization, spatial versus nonspatial visualization, and continuous versus discrete data visualization. Describe what you feel are the aspects of the techniques and data discussed in this chapter that seem to be shared. Are there any techniques that don’t fit this model as well as others?
  3. A new topic called visual analytics is now further differentiating the field. Its definition is “Visual analytics is the science of analytical reasoning facilitated by interactive visual interfaces” [71]. Describe what you feel are aspects that this new field includes which so far have not been discussed in the chapter.
  4. In flow simulation, one often computes a number of different attributes at each time slice and location. Describe at least three distinct ways you could map temperature, pressure, and velocity in a three-dimensional flow field. For each, discuss the accuracy/resolution at which the viewer could attain the data values. Also, discuss the issue of occlusion and the potential for misinterpretation due to partially visible data points.

Chapter 6: Visualization Techniques for Geospatial Data

  1. Map projections are used to visualize geospatial data. Why are these projections difficult?
  2. Different projection techniques do exist. At which points do the different projections have no distortion?
  3. In the R-project distribution (a public-domain statistics tool at http://www.r-project.org) there is a data set called “quakes.” Plot this data and interpret the results of the visualization. Where are the two clear planes of seismic activity?
  4. Shape files represent polygonal boundaries of regions. Look up their definitions, history, and identify issues that would present themselves in coding maps.
  5. Discuss the projection issues in comparing two regions of the world such as Germany and the United States, or even closer ones such as Atlanta and Boston. Note that this issues help explain the distortions we see in a flatland world map.
  6. Use the software Weave to explore various measures and indicators data around the US or World. See http://www.openindicators.org for how to use the software.

Chapter 7: Visualization Techniques for Multivariate Data

  1. For each of the following plot types, describe at least one situation where you would choose this plot over the others.
    • Line plot
    • Area plot (area under line is filled)
    • Bar graph
  2. Rank the techniques presented in this chapter in order of their ability to effectively display data sets with large numbers of records. Write a brief rationale for your choices.
  3. Rank the techniques presented in this chapter in order of their ability to effectively display data sets with large numbers of dimensions/variables. Write a brief rationale for your choices.
  4. Rank the techniques presented in this chapter in order of their ability to convey pairwise correlations between dimensions. Write a brief rationale for your choices.
  5. Compare the asymptotic upper bounds for parallel coordinates, RadViz, and multidimensional scaling using the “big-O” notation.
  6. Display the cars (1993) data set using dense pixel displays.
  7. Display the cars (1993) data set using glyphs.
  8. Display the cars (1983) data using RadViz.
  9. Show how circle segments can be thought of a transformation of parallel coordinates.
  10. Prove that in a line in n-dimensional data maps to a line or a point in a RadViz display. Thus lines are invariant in the RadViz transformation. Show that this also applies no matter what the initial positions of the dimensional anchors are (whether on the circumference or even in a grid).

Chapter 8: Visualization Techniques for Trees, Graphs, and Networks

  1. Give some examples of how rules for graph drawing can conflict with each other.
  2. Compare rectilinear and radial space-filling tree visualization techniques. Under what conditions, or for what tasks, is one better or worse than the other?
  3. Compare node-link and matrix graph visualization techniques. Under what conditions, or for what tasks, is one better or worse than the other?
  4. What is the smallest node-link graph (e.g., smallest number of nodes and links) that you can devise that is nonplanar?

Chapter 9: Text and Document Visualization

  1. Give examples of the suggested computations required for document analysis for the following applications:
    1. identifying plagiarism
    2. determining papers that discuss a specific topic
    3. selecting a Chinese restaurant with good reviews
    4. any other of your choosing
  2. What are some advantages and disadvantages of tag clouds?
  3. Select a document of your choice and generate a tag cloud.
  4. Perform a web search looking for repositories of publicly available text corpora. Retrieve two or three and analyze them in terms of what problems they could be used to solve. What format are they in? What preprocessing is necessary to implement the visualizations given in this chapter?
  5. Repeat the above process, using a newspaper as your source. What sorts of data can you extract from the newspaper? What are the data types? What data sets could you derive by processing the information in the newspaper? Try to design at least one data set for each section of the newspaper.
  6. The techniques in this chapter could be used with television news. How?
  7. Look up Rapid Miner (http://rapid-i.com/) whose core is open source and provide a brief review of its history. RapidMiner can be used as a computational engine for preprocessing text for visualization.

Chapter 10: Interaction Concepts

  1. Give three examples of distortions in two distinct spaces generating identical or very similar results.
  2. Give an example of two specific distortions in different spaces that are commutative, i.e., the results do not depend on the order of application. Give an example where they are not commutative. Are there any general rules you can think of for identifying the conditions under which commutativity would hold or not hold?
  3. In most situations, the user should be able to control the degree of distortion being applied. However, the initial amount should be set to some default level. Discuss how one might set defaults for different kinds of distortion. Consider techniques that are driven by characteristics of the data as well as those independent of the data.
  4. Related to the question above, discuss strategies to set the default extents for different kinds of distortion.
  5. Give examples of distortions with 0th-, 1st-, and 2nd-order continuity. For what reasons might the user choose a particular continuity level?
  6. Select a visualization tool with which you are familiar and examine the types of interaction it supports. List the interaction operators and operands, as well as the parameters of the interaction that the user can control.
  7. Continuing the previous exercise, identify some interaction operators and operands not present in the tool that you feel would be useful additions to the system. Give an example of how they might be used.

Chapter 11: Interaction Techniques

  1. Describe the spaces and interaction techniques in which you feel the fisheye lens algorithm could be effectively applied.
  2. Given a census data set, describe three or more ways you might order the dimensions prior to visualization. What are the strengths and weaknesses of each? You may use the US County Census data set available on the book web site or at the http://www.openindicators.org web page.
  3. When animating a given change, the number of frames over which the change takes place can have a significant impact on the user’s comprehension and satisfaction. Describe the problems that can occur when changes are too fast or too slow, and describe some of the criteria you would use for automatically determining the duration of the animation.
  4. Describe two or more distinct options for animating the shifting of focus on a perspective wall display (Hint: just changing the speed is not sufficiently distinct). Indicate what you feel are the strengths and weaknesses of each.

Chapter 12: Designing Effective Visualizations

  1. Identify at least three problems with the visualization shown in Figure 12.22.
  2. For each of the visualizations in Figure 12.14, suggest at least three modifications that would improve their effectiveness.
  3. Describe four examples of how some of the rules of this chapter may conflict with each other.
  4. Assume that you are plotting the exchange rates for 20 different countries. List at least three ways of ordering the names of the countries and describe why each might be useful.
  5. Other than the figures used in the exercises, find at least three examples of figures in this book that could be improved using design guidelines described in this chapter. Send suggestions for improvements to the authors (yeah, we can take the criticism!).

Chapter 13: Comparing and Evaluating Visualization Techniques

  1. Make a table listing the pros and cons of various evaluation strategies for visualization tools (you may need to read some of the recommended literature first). Are there any strategies that are complementary (e.g., the pros of one address the cons of the other)? This might indicate pairings of strategies that together can paint a clearer picture of the effectiveness of a technique, as compared to running only a single type of evaluation.
  2. Design a set of experiments for evaluating one characteristic of volume visualization techniques. Be careful to specify in detail the task, data, and user characteristics (if human subjects would be involved) that you would be using for the analysis.
  3. Repeat the process using a different task.
  4. Repeat the process using a different characteristic of the technique.
  5. Repeat the process for 2D flow visualization techniques.
  6. Choose a paper from the literature that describes a new visualization technique. Write a summary of what (if any) assessment was performed on the technique, and suggest ways in which further assessment could be performed (you will find that only a small percentage of visualization papers report extensive evaluation).
  7. Skim the papers from one information visualization conference and count how many include evaluations. For each one that does include evaluation, identify the type of evaluation performed (e.g., usability test, expert review, field test, case/use study). Which method(s) appear to be most common?
  8. In light of the above look up the VAST Challenge summary papers (http://vac.nist.gov/index.html) and the participants’ submissions, determine how the visualizations and even the evaluations could be improved.

Chapter 14: Visualization Systems

  1. Examine the functionality of two visualization tools that focus on the same type of data, one commercial and one public domain. How would you characterize their major similarities and differences? Under what conditions, and for what reasons, would you choose to use one over the other (besides, of course, the cost)?
  2. Search the web for a visualization tool not mentioned in this chapter. Write a summary of the system in a style similar to those presented here. Include links to appropriate web pages and published papers. If you’d like, submit the resulting work to the book web site. Those deemed accurate and well written will be posted for others to read.
  3. Choose one of the visualization systems described in this chapter and describe some possible applications for the system. You are encouraged to use the web to identify instances of “real” applications, as well as to use your imagination.

Chapter 15: Research Directions in Visualization

  1. Think of three activities you perform on a daily or weekly basis that you currently do without visualization, but that could potentially benefit from the introduction of visual tools. Describe some ideas for displaying the data or information and for interacting with the resulting views.
  2. Which area of graphics hardware development do you feel will have the biggest impact on the field of visualization: display technology (i.e., bigger screens, more pixels) or rendering technology (e.g., faster GPUs)? Explain your answer.
  3. Give an example of an unstructured data type, and describe what aspects of the data could be visualized.
  4. For at least three different types of data (e.g., spatial, multivariate, relational), discuss the impact on typical visualization techniques if the data is dynamic, rather than static.
  5. Watch a weather report on television on two separate occasions, once with only the sound, and the other time with only the visuals (no text, either). Describe the quantitative and qualitative information you got out of each report, including the strengths and weaknesses of each technique. How much more information would need to be in the visual presentation to equal the quantitative accuracy of the spoken/written report? How much more information would need to be in the spoken/written report to convey qualitative information seen in the visual presentation? Suggest possible enhancements to each.
  6. Choose an application area that currently or potentially uses visualization, and search the web for a published research agenda or list of major unsolved problems. Is visualization mentioned? If so, for what tasks (i.e., exploration, confirmation, presentation)? Try to identify potential uses that were not mentioned.
  7. Talk to a friend, colleague, or family member who you feel is an expert in some area other than visualization. Ask them if and how visualization plays a part in carrying out his or her work. Discuss some potential new ways that person might use visualization. Who knows? It may be the start of a beautiful collaboration!
  8. What user interface issues come into play for visualization for the masses?