Grant: Engage Grants Program

The Web has become the underlying fabric for information dissemination in the modern society. Over the years, we observe two emerging trends: First is that the Web is still mainly for human audience. Much of development of the Web technology focuses on the improvement of content consumption by human readers. In the second trend we witness that there is an increasing amount of content on the Web which are machine generated. To improve the ease of consumption of the machine generated data, the final Web content takes on the form of text based articles, graphical visualization and charts, and tabulation of aggregates. This proposal focuses on algorithms and methods that can assist with the generation of Web content for the audience from structured data sets. In particular, we are interested at content transformation techniques such that the end content is tailored some specific audience. Currently, the standard publishing method requires an analyst to summarize the data set, and manually publish the analysis result in a mixture of text, graphs and tables. Due to the manual intensive involvement of the author, in most cases, only one version of the final publication is generated. This limits the readability and accessibility of the underlying information in several ways. The industry partner (Search Engine People) can benefit from the proposed research by the following improvement in their services: (i) their site content will be targeted to individual users. This leads to higher conversion rate and retention of web traffic; (ii) the visibility of the web content will be improved with more referencing from external sites lead to higher traffic and better Google PageRank; and (iii) better selling point to the company's core business and services.