The PhotoArcs Visualization

Description

The PhotoArcs visualization aims to leverage the convenience, flexibility, and fun of face-to-face sharing along with the persistence and easy remote sharing of online interfaces to encourage users to create digital photo-narratives. These photo-narratives take the form of timelines or “arcs” to take advantage of the strong chronology of photographs and further encourage narrative creation and interaction on a photo-by-photo and story-by-story basis, similar to the interactions in face-to-face sharing. Photos that are re-used in different narratives appear at the intersections of the respective arcs in the interface (Figure 6), allowing the storyteller or the listener to follow tangents of the main story and thus introducing a nonlinear element into the narratives.


Figure 6. Two intersecting photo-arcs with some expanded and some collapsed textboxes

Much of the PhotoArcs interaction is drag-and-drop. Users can click and drag photos around the screen, drop photos on an arc, “tear” photos from an arc (Figure 7), and rearrange photos. Dragging a photo to an arc automatically pops up a text box, into which the user can type. These text boxes by default expand to fit all of their text, but can be resized by dragging their corners or collapsed by clicking on their associated chevrons (« or »; see Figure 6). This allows viewers to peruse just the images in a PhotoArc if they wish. These interaction styles encourage direct manipulation of the narratives and minimize dialogs and other indirect input in order to support more “playful” interaction.


Figure 7. Mock-ups of selecting and removing a photo and its associated text box by dragging

We considered several options for PhotoArcs with so many photos that they cannot fit legibly on the screen, including a telescoping arc, an arc that curls at the ends, and a scrolling arc. Based on feedback from our usability studies, we decided that the best option was an arc that scrolls off the screen and re-centers on a clicked photo.

We also decided to provide several overviews of all arcs to help different navigation styles. Intersecting or related arcs are shown on the screen with the primary arc (Figure 6). A box in the upper-right corner provides an overview of all arcs that a particular user has created (Figure 10). Finally, arcs are loosely arranged according to the time the photographs on them were taken, and the user can zoom out to see all arcs arranged on a scrollable timeline view (Figure 8). In this view, the active arc is still highlighted in the same ways shown above. Hovering over an arc shows the arc’s name and any other metadata the system has stored about the arc.


Figure 8. Timeline overview of PhotoArcs (the mouse is hovering over the active arc, bringing up its name)

In addition to this summary view, a detailed view featuring just one photograph is provided when users double-click a picture (Figure 9).


Figure 9. Detail view of one photograph

Though initially we also wanted to implement a photo-organization tool to help users browse through photographs that are currently not on arcs, we observed that such an effort would, at best, duplicate many other photo-organization research studies such as the information visualization project Amp. Ideally, we would integrate a photo-organization tool such as Amp into the background of the PhotoArcs interface to allow for easy image browsing and selection. For the purposes of this project, though, we decided to focus on the visualization and interaction techniques involving the PhotoArcs themselves, and leave the integration of non-arc photo-browsing to future work. Some screenshots below show clusters of photographs in the background, but we do not discuss the visual properties or implementation of these clusters.

Target Audience

The target audience for PhotoArcs is anyone interested in sharing digital photographs. The best existing fit for this are people who currently take a lot of digital photographs, including photobloggers, family photographers, and other life-chroniclers. In turn, their audience would be their existing audiences for photographs: others at the same event, friends who “should have been there,” distant friends and family, and even strangers.

Implementation

Our initial prototypes consisted of sketches (Figure 1) and low-fidelity prototypes (Figure 12). We used the latter in our user studies, described below. We mocked up additional prototypes using Adobe Photoshop and Adobe Illustrator [2]. We began implementing our interface in Java using the prefuse package [16], particularly experimenting with the force-directed layout functions. Though we did not progress as far in our implementation as we would have liked, we were able to get a sense of the interactions that are feasible to implement and hope to continue the implementation in the future. (Figures 6, 7, and 8 are augmented screenshots from our Java implementation.)

Data Used

Our data consisted of our personal collection of photographs plus those volunteered by the five people who volunteered for our usability studies. We used thumbnails of these photos for the usability studies and for our prototype.

Visual Properties

We leveraged the following visual properties in our visualization to facilitate the user’s creation of online narratives with photos.

Retinal visual properties


Figure 10. An illustration of some of the retinal visual properties employed by PhotoArcs

Color

Position Length, Width, and Size Intersection Direction of motion (for adding or removing pictures) Animation

Gestalt Properties

We also used the following Gestalt properties in the PhotoArcs interface.


Figure 11. An illustration of the Gestalt properties employed by PhotoArcs

Proximity

Similarity Connectedness Grouping Symmetry