The topic of trust and its influence on consumer behavior has received growing attention of late with respect to electronic commerce, especially with regard to establishing and evaluating the trustworthiness of a single web site. However, consumers who rely on the Internet for information gathering are better served by tools that integrate information from many different sources and measure the trustworthiness of the integrated body of information. We propose to develop a methodology that addresses this problem of assembling an information product from many sources and evaluating its trustworthiness. Drawing on techniques in text analysis, knowledge acquisition, graph theory and visualization, and statistical inference, we will enable the user to generate an automated summary from a group of web pages, evaluate its trustworthiness, and visually navigate the information models. The final deliverable is a software package that implements the methodology. The proposed research will benefit the Internet community by providing a new technology that fills the void left by current information retrieval and trust assessment technologies. In addition, the techniques we propose to develop have the potential to contribute to a wide variety of fields confronted with the problem of analyzing samples of graph-valued data.