GAV Flash Tools
The Geovisual Analytics Visualization or GAV component-sharing toolkit and application-building environment is based on the principles behind the Visual Analytics research program including aspects of explorative and collaborative visual data analysis. It contains a collection of visual components, data analysis algorithms, tools that connect the components to each other and data providers that can load data from various sources. The framework is fully integrated with Adobe’s Flex Framework.
GAV Flash uses the Adobe Flash API for graphics and integrates with the Adobe Flex Framework
First developed for interacting with large data using Microsoft’s .Net and DirectX (http://vita.itn.liu.se/gav) , the GAV Flash have now been adapted for the Internet using Adobes Flash V10 basic graphics and Flex 3 for user interface design. GAV Flash is designed with the intention to significantly shorten the time and effort needed to develop sophisticated and dynamic Web-enabled Geovisual Analytics applications through a layered component architecture based on object-oriented class libraries programmed in Adobe’s ActionScript language. A GAV Flash component incorporates versatile interaction methods drawn from many data visualization research areas.
GAV Flash addresses the following generic Geovisual Analytics tasks:
- Shorten development time by utilising already developed and assessed components;
- Design based on cognitive and perceptual principles;
- Appropriate for dynamic multiple-linked views; Visual space-time and multivariate exploration tools;
- Component-embedded interactions including brush, pick, highlight, filter, range slider, view coordination and integrated statics methods;
- Easy to integrate external Flash/Flex components;
- A data cube model for spatial-temporal and multivariate attribute data exploration;
- Integrated mechanism for saving, sharing and packaging the results of a Geovisual Analytics reasoning process;
Below is a list of popular GAV Flash components applied in, for example, eXplorer applications:
Layered Component Toolkit Architecture
Figure:GAV Flash layered component architecture
A keystone in the GAV Flash toolkit is the layered component way of thinking. Instead of making large applications that try to do everything, GAV Flash’s programming practice is to divide a project into as many individually small functional parts as reasonable, and then wrap these within a larger structure. Within a framework such as GAV Flash, the component based structure is a natural way to divide the code base, and using low-level (atomic) layers to assemble higher-level components or booklets is an extension of this method. A GAV Flash application is based on individual functional components, each one performing a specific task, put together in order to create efficient tailor-made applications for any explorative data analysis disaster and emergency work. This layered approach provides a structure for the creation of enhanced or customized versions of existing visualization techniques, encouraging new ideas to be assessed together with standard features without having to rewrite a complete functional visualization component from scratch. Another advantage is that the effectiveness and performance for each part increases. This is due to the simplification in relationships between subparts, as each part in the general case has no need for knowledge about the others which leads to simpler debugging and revision. The component way of thinking enables a short development time, scalability, extensibility, reusability and robustness of components. The applications developed for NCVA projects are based on a low-level component layer (see figure) identified according to their nature in the context of object-oriented programming.
Figure: GAV Flash functional component architecture. All high-level components and booklets rest on the same basis including visualizations, managers and data providers.
All high level components use the data storage class, Data Cube. This class is also used in the analysis to calculate different properties of the indicators, such as percentiles and histograms. Data loading is implemented separately for each data source type. For example, the Excel Reader loads an Excel spreadsheet and creates a data cube containing the data from the spreadsheet and the rest of the framework does not need to know anything about the data source. Data providers can be customized to support many types of sources, Excel is one of the simpler ones while direct data base connections such as SDMX are more advanced and often requires more tasks from the user.
A minimal GAV Flash application contains a data provider and an interactive visualization component such as the Parallel Coordinates Plot (see figure), Choropleth Map, Scatter Plot etc). An application with a single visualization component does not need any selection or visibility managers since these only deals with interaction between two or more components. The visualization components use a Data Cube, supplied by the Data Provider to calculate and create its own special view of the data, such as a colour scheme, correlation plot or a histogram.
Complex and collaborative geovisual analytics sense-making tasks require the external representation and visual organization of information. These methods could help sense-makers compare, organize, comprehend and reflect on what they know, to quickly access specific information when needed, to remember relevant thoughts and ideas, as well as to exchange knowledge and develop a shared understanding with other people. Computer generated information visualizations usually explicitly state relationships among information items thus allowing for quick and non-ambiguous explorations of an information space. Human generated information arrangements are often vague in regards to relationships thus inviting more creative interpretations of an information space. The GAV Flash Framework integrates tools for both collaborative interactive visualization and sense-making. A story indicates a successful suggestion and subsequently fosters additional suggestions based on similar considerations. This learning mechanism allows our storytelling system to improve the accuracy of its suggestions as well as to dynamically adapt to particular users, tasks and circumstances. Colleagues can review a story arrangement and respond with suggestions and comments and subsequently fosters additional suggestions based on similar considerations.
The eXplorer applications (OECD, Factbook, Eurostat, SCB ) facilitate the architecture to support means of capture, add descriptive text, save, packaging and sharing the discovery and results of a geovisual analytics process in a series of snapshots “Story” (image X). When the button “Capture” in the Story Editor is pressed, the state of each GAV Flash view in eXplorer is saved together with user-defined metatext. Before closing the application, the user exports the story into a XML formatted file. Team members can through descriptive text combined with interactive visualization follow the analyst’s way of logical reasoning by loading selected stories. At any time a team member can access stories and apply them in eXplorer or any other GAV Flash application assembled from the same component. A comprehensive story in the context of a remote collaborative sense-making activity can thus be created by the analyst through a set of linked snapshots (chapters). Users will discuss relevant issues through storytelling based on solid evidence, thus raising awareness and increasing the common knowledge on a certain phenomenon.
Last updated: 2010-01-29