GoOLAP: Search. Analyze. Decide

GoOLAP is a project of the Technische Universität Berlin. The project took place under supervision of Alexander Löser and Volker Markl.It resembles very much the idea of an entity-centric tool for the aggregation of data. Here is how it is described on the Web site:

The World Wide Web is today one of the largest databases. Lots of information is provided as unstructured text in e.g. news feeds. Most of the content is user generated content and hardly to aggregate or control. Today's search engines are able to find keywords, but the user still has to review the data sources to find the information he searches for. The main idea of the "GOOLAP" project is to recognize the search intention of a user and to provide the information in real time from an aggregation of web data sources of different kinds. We focus on simple intentions like "show me all information about the entity 'Barack Obama'" or list members of the class "president". For solving these queries our GOOLAP.info prototype addresses three main problems: Extract entities and relations from a large corpus of a news-data, recognize the search intentions from the user and display aggregated extracted data in a simple, fast and intuitive fashion.