SIGIR 2007 Proceedings Demonstration Search Results using Timeline Visualizations Omar Alonso Dept. of Computer Science University of California, Davis Michael Ger tz Dept. of Computer Science University of California, Davis Ricardo Baeza-Yates Yahoo! Research Barcelona, Spain oralonso@ucdavis.edu ger tz@cs.ucdavis.edu rbaeza@dcc.uchile.cl Categories and Subject Descriptors: H.3.1 [Information Storage and Retrieval]: Content Analysis and Indexing - Linguistic processing. H.5.2. [Information Interfaces]: User Interfaces. General Terms: Design, Experimentation. Keywords: information visualization, temporal information 1. OVERVIEW As search applications keep gathering new and diverse information sources, presenting relevant information anchored in time becomes more important. Time can help in describing the context of a document or document collection or in recreating a particular historical period. Temporal information is available in every document either explicitly or implicitly. Recognizing such temporal information and exploiting it for document retrieval and presentation purposes are important features that can significantly improve the functionality of today's search applications. A look at the functionality of any of the current search engines shows that temporal aspects of documents are exclusively used to sort the hit list by date, which is primarily the date a Web page or a document has been created or last modified. In this demonstration, we present a new approach in which search results are arranged in a well-defined timeline using a few existing visualization metaphors. To take advantage of time related information for IR purposes, we process each document in a document annotation pipeline that extracts temporal expressions and produces output in TimeML. Recently TimeML has emerged as a standard markup language for events and temporal expressions in natural language [1]. As the basis for anchoring documents in time, we assume a discrete representation of time based on the Gregorian Calendar, with a single day being an atomic time interval called chronon. Our base timeline, denoted Td , is an interval of consecutive day chronons. For example, the sequence "March 12, 2002, March 13, 2002, March 14, 2002" is a contiguous subsequence of chronons in Td . For a query, the search engine returns an intermediate representation of the hit list grouped by year. This is later transformed to a particular visualization format. The timeline construction is based on a clustering algorithm that uses temporal expressions (chronons), which have been extracted from the documents during the above annotation process and anchored in a timeline. For our prototype, we use the TimeBank corpus, which contains news articles that have been annotated using TimeML. In the first visualization, we use Simile TimeLine [2] to construct a timeline that consists of two bands representing different time scales: decade and year. Both bands are synchronized such that Copyright is held by the author/owner(s). SIGIR'07, July 23­27, 2007, Amsterdam, The Netherlands. ACM 978-1-59593-597-7/07/0007. Figure 1: Search Results with SIMILE panning one band also scrolls the other. In the second visualization, we use Inxight's TimeWall [3] to define cards with attributes that allow their placement in the wall along the timeline. In both metaphors the user can click a particular document and the content of the document is shown in a separate frame. Both approaches not only demonstrate the high utility of temporal information associated with search results and documents, but they also indicate an important direction in the development of future search engine features. Figure 2: Search Results with TimeWall 2. REFERENCES [1] TimeML 1.2.1 Specification: http://www.timeml.org [2] TimeLine: http://simile.mit.edu/timeline [3] TimeWall: http://www.inxight.com/products/sdks/tw 908