Visual Analytics Dashboard
Visual analytics combines data science and visualization methods to provide insights that support decision making and strategic planning. Powered by the latest text mining and artificial intelligence algorithms, the webLyzard dashboard is a prime example for this quickly emerging field. It serves as a versatile content exploration tool for Web intelligence applications.
The ability of the system to analyze the public debate in real time enables you to incorporate the latest trends into your decision making. The video tutorial embedded on the right not only explains the usage of the dashboard, but also presents various use cases. You can forecast future topics, for example, or identify opinion leaders who might help to amplify your message. Impact metrics and emotion detection algorithms also determine how stakeholders perceive your organisation’s activities in a given context. This context is usually multidimensional in nature. Accordingly, the dashboard helps to track and visualize emerging stories across sources, regions and languages.
The screenshot below summarizes the various views of the dashboard to access and navigate the content repository. A description of these views follows below. Whenever you interact with the system and use the interactive features of one particular view, your actions will trigger an immediate update of all other views.
The text input box in the upper left corner is the currently active context filter (“*” indicates that the filter is turned off). The values indicate the number of matching documents, their color-coding reflects average sentiment. If the rectangular marker on the left is selected, results matching the context filter are shown in the trend chart as a dotted line. On hover, the gear icon provides access to advanced filter options.
The main menus of the visual dashboard include configuration, keyword filter and data export in various formats, e.g. PDF reports in A4 or 16:9. Two drop-downs allow selecting the date range and data sources (news media, social media, etc.) to be used for the analysis. The area chart previews the results matching the context filter over a longer time span. The time interval is a global setting that not only affects the trend chart, but also the content area and all visualizations. You can specify the interval by choosing a “from” and “to” date, or by selecting the past [n] days, weeks or months. Below, the info bar either shows context-aware help texts or a list of the most frequent sources or referenced entities (persons, organizations or locations).
Left Sidebar – Default Settings
- Associations. This sidebar element displays a list keywords associated with the context filter, in light of the selected source(s) and time interval. It is the degree of association that determines the sequence of terms. You have the following options:
- Click on a label to access the tooltip. You can either trigger (Replace) or refine (Restrict, Extend) a search, or Exclude certain terms from the search results. This tooltip is available across various dashboard features, e.g. the visual analytics tools listed below. It includes a small line chart preview and list of associated keywords. Moreover, you can combine multiple searches or filtering results by metadata attribute, e.g. positive sentiment;
- Activate checkboxes to include the certain topics or metadata attributes in the search. The bar chart icon visible on hover left of the “Associations” header either selects the first five items, or deselects all items.
- Metadata. Below the associations, the left sidebar provides access to the metadata view that segments results by source, sentiment, recency, language or country.
- Bookmarks. This section helps to create and manage topics. On hover, a gear icon lets you view, configure, rename or delete bookmarks. In addition, you can specify the sort order and set email alerts.
Left Sidebar – Optional Elements
- Search History is not shown in the default setting, but can be activated via the Configuration menu. It is a chronologically sorted list of recent searches, showing either the search term or a bookmark label.
- Clipboard is an additional widget activated automatically. It can appear when the context filter changes while checkboxes of Associations are active. Otherwise, changing the content filter could make them disappear. In such situations, the Clipboard sidebar ensures continued access to these terms.
- Line, Donut and Bar Charts. The interactive trend chart shows (i) the share of voice, a comparative measure of attention based on the relative number of mentions, (ii) the frequency of selected topics in the specified time interval, (iii) the average sentiment regarding these topics, or (iv) the level of disagreement, measuring how strongly the sentiment fluctuates. The default moving average interval for the data analysis is seven days. Selecting longer intervals dampens the impact of short-term fluctuations and thereby highlights longer-term trends.
Main Content Area
- Search Results and Cross-Media Analysis. The section below the trend chart offers five main categories to explore the search results: Documents, Sentences, Sources, Entities and Relations. These categories offer multiple representations, typically at least one list view and one visualization. You can select a word tree for sentences, for example, or scatter plots for entities and sources.
- Interactive features in the ‘document’ and ‘sentence’ views include: (i) mouse-over document preview; (ii) initial click to show an extended quote; (iii) second click to display the full text and various annotations including keywords, sentiment, content source and referenced location.
Right Sidebar – Visual Analytics Tools
- Maps and Visual Analytics Tools. Interactive tools help to explore the underlying knowledge base. To investigate topic associations, the dashboard provides a tag cloud (alphabetical) and a keyword graph (hierarchical). Additional visualizations include a geographic map to render the regional distribution of search results and the cluster map to show emerging stories (= sets of related documents).
- Interactive Widget Controls. You can select the visualizations via the “Configuration” menu. Via drag-and-drop you can reposition them in line with your current task. The ‘pause’ button in the upper right corner disables dynamic updates. This includes the automated zoom in the geographic map, for example, or the node positioning in the keyword graph. Clicking on the ‘maximize’ button increases the size of the window and adapts its content.
Visual Analytics Publications
- Scharl, A., Herring, D., Rafelsberger, W., Hubmann-Haidvogel, A., Kamolov, R., Fischl, D., Föls, M. and Weichselbraun, A. (2017). Semantic Systems and Visual Tools to Support Environmental Communication”, IEEE Systems Journal, 11 (2). pp. 762-771.
- Scharl, A., Hubmann-Haidvogel, A., Sabou, M., Weichselbraun, A. and Lang, H.-P. (2013). From Web Intelligence to Knowledge Co-Creation – A Platform to Analyze and Support Stakeholder Communication, IEEE Internet Computing, 17(5): 21-29.
- Hubmann-Haidvogel, A., Scharl, A. and Weichselbraun, A. (2009). Multiple Coordinated Views for Searching and Navigating Web Content Repositories, Information Sciences, 179(12): 1813-1821.