search thumbnail

Search and Content Retrieval

The main search bar is located at the top of the page. Search terms added here trigger a general, full-text query over the entire collection of documents. Each query returns a list of relevant documents while also updating various context views including the tag cloud, keyword graph, and geographic map.

The search bar features a drop down menu for auto-completion, returning suggested keywords and entities. The system also supports typical wildcard characters, such as the asterisk (*) for representing any number of unknown characters and the question mark (?) for representing exactly one character. More specific search options including Boolean operators for source, location, date and other metadata elements are described in the advanced search section.

The tabs of the search results section shown in the following screenshot let users select the most appropriate representation for their current task – either textual or visual, on various levels of aggregation:


Content Categories

Documents. Once a user has entered a search query, the system lists all documents that match this query. The results can be sorted and organized using the dates, sentiment and sources columns. Note that only a subset of the documents is displayed in the list. Typically these are the 50 most relevant documents to the search query, while the total number of matching documents can be seen at the top of the page – e.g. 0 – 50 out of 4866. Clicking on the text block extends the quote, a second click activates the full-text mode.

  • Full Text View. When the full text of a document is shown, the header of the page includes document keywords and the URL of the source, while the footer summarizes the document’s other annotations including source category, source location, target location, sentiment, and relevance. Clicking on the sentiment and relevance values activates a slider to suggest alternative values.
  • Metadata View. This tab reveals how the sentiment of a document has been computed by highlighting terms that are contained in the sentiment lexicon (green = positive, red = negative), as well as negation triggers (blue) that reverse the polarity of a sentence.

Sentences. Alternatively, the system lists matching quotes and groups them by document. Users can sort the results by source, date of publication, and the sentiment on both the document level and the sentence level (the color of the term reflects sentence-level sentiment).

Word Tree. An interactive tool for showing the lexical context of search terms.

Entities. This tab displays a table of locations, persons and organizations referenced in the search results (automatically detected by the Recognyze component). The list can be sorted by entity name, the number of entity references, and the average sentiment towards the entities. Clicking on a row triggers a query for all documents tagged with a specific entity.

Entity Map. An interactive tool for analyzing relations between entities and topics.

Sources, Source Map. Please refer to the cross-media analysis document.

Last major update with release 2013-06.2 (Eastern Fence Lizard).