Bungee View lets you search, browse, and data-mine image collections. It shows an overview of the entire collection, in addition to the familiar result list and item details of other search interfaces.
Welcome to Bungee View |
Foo |
Bungee View is designed to support non-technical users in gaining an understanding of an image collection as a whole, and in finding patterns in their meta-data, while they search and browse. Since this is a research project, your usage will be recorded to help improve Bungee View. We will use your IP address to track your repeat visits, but will not attempt to identify you as a person. Feedback is welcome at (Mark Derthick).
For a brief tour, watch this two minute video as an avi or a worse quality mov (the DivX codec should work for either).
| January 2008 | New version released. 1) Bars now go up for positive tag associations and down for negative associations. 2) There is now support for alphabetic tag selection by clicking, dragging, or typing (expert mode only). |
| November 2007 | Bungee View wins award at the 2007 Information Visualization Symposium, based on this eight-minute video showing expert-level use on a database of movies. |
| November 2007 | There is now a project page for Bungee View Administrators and Developers. |
* If your browser doesn't know how to open this link, it means you need to install Java 1.4 or later, which includes the Java Web Start plug-in.
Bungee View lets you search, browse, and data-mine image
collections. It shows an overview of the entire collection, in addition to the
familiar result list and item details of other search interfaces.
For each metadata category, vertical bars show tags associated with your query. In the screenshot above, the query is for images that show Places in Asia. The bottom row of bars shows tags for the category Kinds of People. Wide bars represent Kinds of People that appear in many images in the collection. Tall bars represent Kinds of People associated with your query, either positively if the bar goes up, or negatively if the bar goes down. For instance, images that show Educators & Communicators are six times as likely to show Places in Asia as images that show other Kinds of People, while those that show Manual Laborers are only one tenth as likely. The vertical scale ranges from 1/100th to 100 times the likelihood of other tags in the category, and is non-linear.
The bar color indicates whether the association is statistically significant. Dull green bars show significant positive associations. Dull red bars show significant negative associations. Gray bars show associations that are not significant. Both the Communicators & Educators association and the Manual Laborers association are significant. Saturated green and red bars (such as Asia) are for values that are part of the query.
Web search engines have attracted widespread demand for information retrieval from unstructured documents. The number of structured and semi-structured documents available on the Web is also huge, and collections of these are more amenable to data mining. Yet there has been no similar explosion of interest in this kind of exploration. Finding patterns in databases of political contributions, environmental data, or hospital and school performance would surely interest many citizens. The main research question for this project is how to support such exploration for users with little or no training in statistics or programming. In contrast to other data-mining systems, Bungee View focuses on learnability, responsiveness, robustness, and providing a satisfying user experience.
Bungee View version 11/2007, Copyright (C) 2007 Mark Derthick
Bungee View comes with ABSOLUTELY NO WARRANTY; This is free software,
and you are welcome to redistribute it under certain conditions; Choose
"About Bungee View" on the Help menu for details.