The visual information exploration and retrieval interface of the Media Watch on Climate Change builds upon a portfolio of annotation services. These services create a contextualized information space by enriching the content repository with geospatial, semantic and temporal annotations, and by applying concepts embedded in a climate change ontology as a controlled vocabulary for structuring the stored information. The annotations provided by the semantic services describe complex relations. To visualize these relations, the portal integrates geographic displays, domain ontologies, tag clouds, maps of document clusters, and just-in-time information retrieval agents suggesting documents about similar topics and nearby locations. Geospatial technologies are crucial for this integration – virtual globes are not only a natural way to show the most relevant documents in their regional context, but also excellent platforms to deliver interactive visualizations of non-geographic data. Such visualizations help users to perceive and understand context when navigating large archives of Web content, and to analyze the perceptions and informational needs of various stakeholders.

Table of Contents

1. Interface Structure
2. Trend Charts
3. Just-in-time Information Retrieval Agents
4. Full Text Queries
5. Visualizations

Interface Structure

Drop-down elements in the main menu let users choose a specific date as well as the document source (News Media, Blogs, Eco-NGOs, Fortune 1000). The trend chart in the upper left corner shows keywords and trends in the previous ten weeks. The content view below shows the active document, including its mirror date, place of publication (source geography), and primary location that is being referenced (target geography). Further below, just-in-time information retrieval agents list documents referring to similar topics and nearby locations. Visual components on the right side of the screen show various dimensions of the annotated documents. The information retrieval agents as well as the maps can be repositioned using drag-and-drop operations.

Trend Charts

This window shows the five highest-ranking keywords for a given week, including the observable trends in the preceding ten weeks in terms of attention and sentiment (users can choose between the two categories). Clicking on a keyword in the legend triggers a full-text query; clicking on the colored series marker hides the series from the chart. Since the vertical axis is rescaled automatically, this feature is particularly useful if one keyword dominates the coverage and therefore obscures the distributions of the other four keywords.

Just-in-time Information Retrieval Agents

Two just-in-time information retrieval agents list documents referring to similar topics and nearby locations relative to the Active Document (the term ‘nearby locations’ referring to the five documents with the closest target geography).

Clicking on the related references extends the quoted text, clicking on the circular marker on the left activates that particular document. Back and forward buttons enable the user to browse the list of related documents. The RSS links provide a continuously updated list of related documents (see below for a more detailed description).

Full-text Queries

Simple Search. The most common way of querying for information is the search box, located in the top bar of the portal. Queries return the most relevant documents in the Similar Topics view and display the full text of the highest-ranking one in the Active Document view, including its source and target geography. Once a user has entered a search query, an additional window displays a list of quotes that include the target term as 'concordances' (centering the target term and showing its immediate context in the document). By clicking on the column headers, the quotes can also be sorted by their sentiment (+/-) and date of publication.

Advanced Search. In addition to simple full-text queries, advanced search options allow including document metadata in the search query. The title field of the documents can be queried using the title: operator; searching for “carbon dioxide title:sea” will only return documents which contain “sea” in their title. The site: operator limits to search to a given domain; searching for “carbon dioxide site:cnn.com” returns documents on “carbon dioxide” mirrored from the CNN Web site. Documents in a certain time frame can be searched by using the from: and to: operators; searching for “carbon dioxide from:2009-01-01 to:2009-01-31” will only return documents published in January 2009. Source and target geography of the documents can be considered with the source: and target: operators; searching for “carbon dioxide source:US target:Kyoto” will return documents about carbon dioxide published in the US and referencing Kyoto as the main location in the text.

Visualizations

The maps on the right side facilitate access to the underlying knowledge base. Clicking on the ‘maximize’ button increases the size of the maps; clicking on the ‘popup’ button opens the map in a separate browser window (which allows using the system in multiple-screen configurations). Maps can be rearranged by dragging them to the desired position, and switched on and off using the buttons in the top bar. As an alternative to entering query terms (see previous section), users can click on any position in the maps (not only on the document markers) to retrieve articles related to that particular location, topic or domain concept. The different views are therefore said to be tightly coupled – user actions in one window trigger an immediate update of all other displays.

Semantic Map

The concept of ‘location’ can be implemented in innovative ways that transcend the traditional geographic interpretation. The Semantic Map resembles an ordinary landscape at first sight. Instead of geographic proximity, however, it represents semantic similarity between documents. At the time of map generation, its topography is determined by the content of the knowledge base. The peaks of the virtual landscape indicate abundant coverage on a particular topic, whereas valleys represent sparsely populated parts of the information space. Such information landscapes allow clustering and visualizing massive amounts of textual data.

Ontology

The ontology graph displays a clickable domain ontology that matches documents and concepts to help users determine their current location in the information space. The ontology graph depicts hierarchical relations as arrows. The currently active document is highlighted by a red asterisk, and the letters [1-5] represent the classification of the five highest-ranking documents in the similar locations view. Clicking on a concept activates the top-ranked document for this particular concept. The ontology itself has kindly been provided by the Climate Program Office of the U.S. National Oceanic and Atmospheric Administration (NOAA; www.climate.noaa.gov).

Geographic Map

This view shows the source and target geography of documents. The currently active document is highlighted by a yellow asterisk, and the letters [a-e] represent the five highest-ranking documents in the similar topics view. If users have an interest in a specific location rather than a topic, they can click anywhere on the map to trigger a query for documents referencing this region, instead of using the search form (this principle of fully interactive tightly-coupled views runs across all visualizations).

If users enter a search term, the set of results is visualized in the geographic map as well. Red circular markers show the target geography of the found articles, the diameter of the marker representing the number of matches for a given location. Using a color range from yellow to red, trajectories link the source and target geography of an article (source geography = location of the publisher; target geography = main location referenced in the document). The little ‘+’ symbol on the right side of the window allows users choosing alternative base maps (e.g. NASA Blue Marble, Political Borders, Google Terrain, etc.), as well as deactivating the circular markers or trajectories.

Tag Cloud

The tag cloud visualizes keywords identified in the corpus. Terms are arranged alphabetically, both size and color of the terms are proportional to the keyword’s importance in the corpus (i.e., its relative frequency). The tag cloud displays more frequent terms in black, using a larger font, less frequent terms use a smaller font and are visualized in light gray. Users can select a term by clicking on it, which triggers a full text search and displays a list of matching concordances.

Date Selector

Users can access historic data by selecting a certain point in time using the date selector. Dates can be selected by navigating one day (or week) into the past or future by using the provided back and forward buttons, or by selecting a specific date using the drop-down calendar. Selecting a new date automatically updates the trend charts (displaying the data for the selected week), the active document (matching the top keyword for the selected date), as well as the semantic map in case the new date belongs to a different weekly snapshot.