Extend Excentric Labeling: Automatically adjust focus area ...



Extend Excentric Labeling: Automatically adjust focus area and Show Summary Information

Tao Zhan

Computer Science Department

University of Maryland

College Park, MD 20782, USA

taozhan@cs.umd.edu

ABSTRACT

The widespread use of information visualization is hampered by the lack of effective labeling techniques. A new dynamic labeling technique is proposed by Fekete and Plaisant [Fekete and Plaisant, 1998] called Excentric Labeling. This technique labels a neighborhood of objects located around the cursor. It does not intrude into the existing interaction, is not computationally intensive, and was easily applied to several visualization applications. This paper discusses and implements two extensions for Excentric Labeling. One is to force the number of labeling objects to be a fixed number by adjusting the size of the focus area; The other is to show summary information of the data objects in the focus area by using a bar chart instead of just showing the labels of the objects.

KEYWORDS

Visualization, Label, Dynamic labeling

INTRODUCTION

A major limiting factor to the widespread use of information visualization is the difficulty of labeling information abundant displays. Information visualization uses the powerful human visual abilities to extract meaning from graphical information [Card et al., 1998; Cleveland, 1993]. Color, size, shape position or orientation is mapped to data attributes. This visualization helps users find trends, and spot exceptions or relationships among elements on the display. Experimental studies have been able to show significant task completing time reduction and recall rate improvements when using graphical displays instead of tabular text displays (e.g., [Lindwarm-Alonso et al., 1998.]). However, textual information, in the form of labels, remains critical in identifying elements of the display. Unfortunately, information visualization systems often lack adequate labeling strategies. Often labels are entirely missing and users have to peck at graphical objects one at a time. Sometimes labels overlap each other to the point of obscuring the data and being barely usable; or they are spread out in such a way that the relation between objects and labels becomes ambiguous. The problem becomes acute when the data density increases and the labels are very long.

Fekete and Plaisant propose a new dynamic technique "Excentric Labeling" to label a neighborhood of objects (Figure 1). Excentric Labeling provides labels for a neighborhood of objects. The focus of the labeling is centered on the cursor position. Labels are updated smoothly as the cursor moves over the display, allowing hundreds of labels to be reviewed in a few seconds. The color of the label border matches the object color.

In this paper, I will propose two extensions for this technique. One is automatically adjusting the size of focus area to let the number of displaying objects be a fixed number or around a fixed number. The other is that instead of showing the labels of the objects in the display, some summary information about the objects in the focus area is shown (Figure 2).

[pic]

Figure 1: Excentric Labeling

[pic]

Figure 2: Excentric Labeling showing summary bar chart

I will first describe labeling techniques in general, then describe the Excentric Labeling in detail as well as discuss its benefits and limitations, and illustrate how the two extensions are implemented, and give an example application for showing summary information technique. I then give the conclusion and future work.

LABELING TECHNIQUE

The labeling challenges can be stated as follows: given a set of graphical objects, find a layout to position all names so that each name (label) is:

Readable.

Non-ambiguously related to its graphical object.

Does not hide any other pertinent information.

Complete (the labeling of all objects). This is desired but not always possible.

Labeling techniques can be classified into two categories: static and dynamic. The goal of static labeling is to visually associate labels with a maximum number (hopefully all) of the graphic objects in the best possible manner. Dynamic labeling began with interactive computer graphics and visualization. Two attributes account for the "dynamic" adjective: the set of objects to be labeled can change dynamically, and the number and layout of displayed labels can also change in real time, according to user actions.

Static techniques have been used for a long time in cartography and other general-purpose visualization. Since static technique has to find "the" best possible labeling, the set of objects has to be chosen carefully to avoid too high density in objects or labels. In cartography, this is achieved by aggregating some information and getting rid of others (this process is called "generalization" or “sampling” ). Other techniques such as "No Label" and "Rapid Label-all" lead to multiple overlaps and data occlusion [e.g., in the hyperbolic browser[Lamping et al, 1995]). Also common is the "Label-What-You-Can" technique in which only labels that fit are displayed; other labels that would overlap or occlude data objects are not shown. The main disadvantage of static labeling is that it has to arrange objects and their labels into fixed position as well as prevent ambiguous relations and overlapping. This is often in conflict with the complete requirement and the situation is worse when data objects are dense.

Dynamic labeling techniques are more varied. The classic infotip or "cursor sensitive balloon label" consists of showing the label of an object right next to the object when the cursor passes over it. The label can also be shown on a fixed side window, which is appropriate when labels are very long and structured. Some zooming techniques zoom into large display until labels appear. An overview + detail technique will be effective when objects are separated enough in the detail view to allow labels to fit (not guaranteed). Sampling techniques choose fewer sampling labels to display.

EXCENTRIC LABELING

Excentric Labeling is a dynamic technique of neighborhood labeling for data visualization. When the cursor stays for more than one second over an area where objects are available, all labels in the neighborhood of the cursor are shown without overlap, and aligned to facilitate rapid reading. A circle centered on the position of the cursor defines the neighborhood or focus region. A line connects each label to the corresponding object. The style of the lines matches the object attributes (e.g., color). Once the Excentric labels are displayed, users can move the cursor around the window and the Excentric labels are updated dynamically. Excentric Labeling stops either when an interaction is started (e.g., a mouse click) or the user moves the cursor quickly to leave the focus region.

Excentric Labeling uses three different algorithms to compute the layout of labels: non-crossing lines labeling, vertically coherent labeling and horizontally coherent labeling. Each uses a different method to compute the initial position, to compute the ordering, to assign the labels to the stacks and to join the labels to their related graphic objects. The Non-crossing lines labeling layout does not maintain the vertical or horizontal ordering of labels, but avoids line crossings, while vertical or horizontal layout maintains the relative vertical or horizontal order of labels. Excentric Labeling also can deal with window boundaries by moving the truncated labels from left stack to right stack or vice verse. And it also can reveal objects clumped together on the display. (figure 3 and figure 4)

[pic]

Figure 3: Excentric Labeling reveals objects clumped together.

[pic]

Figure 4: Excentric Labeling handles window boundaries.

The advantage of Excentric Labeling is that it can show a group of labels without overlapping and being aligned to facilitate rapid reading. It can dynamically update according to the movement of the cursor, it does not require the use of special interface tool. Labels are readable, non-ambiguously related to their graphical objects and they don't hide any information inside the user's focus region.

There are also some problems in Excentric Labeling. If object density is very high and too many labels are in certain focus area, it is hard to display them completely. The labels could be too long to fit on either side of the focus point. The position of the same label could change as the cursor moves and lead to discontinuities in the placement of the labels.

AUTO-ADJUST FOCUS

When the data objects in certain focus region is of very high density or the focus region is too large, there could be a large number of labels being displayed. It is hard for the user to tell the correspondence between the labels and the objects. Usually the user might want to control the maximum number of labels that can be shown. Excentric Labeling allows the user to specify a fixed maximum number of labels and a fixed size of circle region for the focus area, but those two parameters are independent. So if there are more labels than the maximum number in this region, what Excentric Labeling will do is to display a random set of maximum number of the labels and display the total number of labels in this focus region. In some cases this might not be what users want since users do not determine the random set of labels. One extension to this problem is allowing the size of the focus region to be dynamically changed to force that the number of displayed labels is around certain desired number.

[pic]

Figure 5: Old Excentric Labeling showing random 20 labels, while the actual number of labels is 35.

As the cursor moves, the size of the circle centered in the cursor position changes to include more data objects or exclude more data objects, so that the total number of displayed labels is around the maximum number user specified. Figure 5 shows old Excentric Labeling display random 20 labels in this region while the actual number is 35. Figure 6 shows that when auto-adjust focus technique is applied, the size automatically shrinks with only 20 data objects left in focus region. Figure 7 shows when you move the cursor to another position, the number of labels displayed is always around 20.

[pic]

Figure 6: Auto-adjust focus technique shrinks the region with only 20 labels left in it.

[pic]

Figure 7: The size of focus region automatically updated to maintain fixed number of labels in the focus region.

SHOW SUMMARY INFORMATION

One key feature of Excentric Labeling is that it always has a certain focus area and a group of objects in this region. So another extension is that instead of only showing the labels of the data objects, we may show some summary information of those group objects. Because of the dynamical feature of Excentric Labeling, this summary information will be updated according to the cursor movement. Users may use this technique to collect some statistics or summary information for different group of data around the window and reviewed them in a very short period of time. This summary or statistics information will avoid the density problem in displaying labels and give useful information to users.

A bar chart is used to show the summary information. One possible type of summary information is the color distribution in this focus region. So as user moves cursor in the graph, a dynamical bar chart will show how the object colors are distributed in this focus area. (Figure 8)

[pic]

Figure 8: Showing summary color distribution bar char for the focus region.

User can use an option window triggered by right clicking the mouse to enable or disable the Auto-Adjust Focus and Show Summary Info option. (These two options are independent and can be combined together). User can also choose the orientation of the bar chart. Figure 9 shows the option window and Figure 10 shows a vertical bar chart.

[pic]

Figure 9: Option window.

[pic]

Figure 10: Vertical bar chart.

Since Excentric Labeling can be easily imported to other applications and the attributes of the data objects can be variations of numeric value, other statistics such as min, max, mean, standard deviation or other value distribution graph could also be shown. Those dynamic summary information will give the user more abilities to analyze the data.

SAMPLE APPLICATION

Showing summary information technique in Excentric Labeling can be applied to many current applications. For example, in cartography area, if users want to find a good place to live, they might be interested in the environment and facilities around this area. They can use showing summary information technique to dynamically view the distribution of certain kind of objects or facilities in some regions, such as lakes, forests, grass, schools, hospitals, and police stations. By moving the cursor around the map and looking for the distributed bar chart, users may find the ideal place to live. Figure 11 shows a sample scene for this application.

[pic]

Figure 11: Excentric Labeling showing the distribution of the environment and facilities in certain area.

CONCLUSION

Despite the numerous techniques found in visualization systems to label the numerous graphical objects of the display, labeling remains a challenging problem for information visualization. Excentric Labeling provides a novel way for users to rapidly explore object descriptions once patterns have been found in the display and effectively extract meaning from information visualization. Auto-adjust focus technique extends Excentric Labeling to avoid too many labels in certain focus region and allow users to set the number of labels displayed to a fixed number. Showing summary information technique displays statistics summary information for the objects in the focus region which user may be interested in. And one possible application in cartography field is shown in this paper for showing summary information technique.

FUTURE WORK

Future work of the extension could be to include more types of chart to display summary information, such as curve, pie chart, history graph and allow users to have more controls to specify the options for those summary charts. One option for the chart graph is to display the graph in a fixed position instead of keeping moving around the graph window. Another possible extension for Excentric Labeling could be combining magic Len with Excentric Labeling to add filter and selection operation to those data objects in focus area.

REFERENCES

1. Fekete, Jean-Daniel and Plaisant, Catherine, Excentric Labeling: Dynamic Neighborhood Labeling for Data Visualization.

2. Robertson, George G. and Mackinlay, Jock D., The Document Lens.

3. Ahlberg, Christopher and Shneiderman, Ben, Visual information seeking: Tight coupling of dynamic query filters with starfield displays, Proc. CHI’94 Conference: Human Factors in Computing Systems, ACM, New York, NY(1994), 313-321 + color plates.

4. Christensen J., Marks J., Shieber S. Labeling Point Features on Map and Diagrams, to appear in Transactions of Graphics.

5. Card, S, Mackinlay, J., and Shneiderman, Ben, Readings in information Visualization: Using Vision to Think, Morgan Kauffman Publishers.

6. Plaisant, Catherine, Rose, Anne, Milash, Brett, Widoff, Seth, and Shneiderman, Ben, LifeLines: Visualizing personal histories, Proc. Of ACM CHI96 Conference: Human Factors in Computing Systems, ACM, New York, NY(1996), 221-227, 518.

7. Cleveland, William, Visualizing Data, Hobart Press, Summit, NJ (1993).

8. Lindwarm D., Rose, A., Plaisant, C., and Norman, K., Viewing personal history records: A comparison of tabular format and graphical presentation using Lifelines, Behavior & Information Technology.

9. Lamping, John, Rao, Ramana, and Pirolli, Peter, A focus + context technique based on hyperbolic geometry for visualizing large hierarchies, Proc. Of ACM CHI’95 Conference: Human Factors in Computing Systems, ACM, New York, NY (1995), 401-408.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download