Uncertainty Visualisation Selector

About

The UVS is a simple web-based interface to a solid usability research action about selected spatio-temporal uncertainty visualisation methods conducted by Hansi Senaratne. These selected methods were priorly categorised according to a set of parameters (supported uncertainty type, data type and data format) that are believed to be the most influential for a successful visualisation. Basing on the hypothesis that usability not only varies between different uncertainty visualisation methods, but also between different user groups who makes use of these methods, an online survey was conducted on 81 participants from different domains to assess which uncertainty visualisation method(s) suited best for each domain user. A poster and its abstract that was done on this research can be found on the links below. A full review of this research can be expected in the near future.

How it Works

A user puts in the specifics of a dataset, namely type of uncertainty, data format, the measurement scale, and domains. The UVS calculates the most suitable visualisation methods and shows an ordered list including additional information for each method.

Development

Source Code

SCM access: https://github.com/GeoViQua/UncertaintyVisualisationSelector

Check out and build project

The UVS uses a Maven (http://maven.apache.org/) build process. Follow these steps to download the source code from the source code management server for building it locally, and running a local test server.

Prerequisites:

• Eclipse, http://eclipse.org/
  • Maven often has problems when running Eclipse with a JRE, not a JDK. Also make sure maven itself is running in a JDK.
• m2e plugin, http://eclipse.org/m2e/
• SVN connector
  • see Help > Install New Software... > Work with: --All Available Sites-- and filter for "svn"
• Java JDK installed, JAVA_HOME configured

1. Start Eclipse
2. Open perspective "SVN Repository Exploring"
   • File > New > Repository location
   • https://svn.52north.org/svn/geostatistics/main/geoviqua/
   • add login (for developers with 52N account only)
   • Click Finish
3. Navigate to directory /uvs/trunk
4. Right click on the directory trunk
   • Check out as Maven Project ...
   • Click Finish
5. You have a new Eclipse project called 52n-uvs
6. Open Run configurations
   • Maven Build
     • Add new run configuration
       • Name: Package
       • Base directory:
         • Click Browse Workspace..., then select project folder
         • ${workspace_loc}/${project_name}
       • Goals: package
       • Details: http://maven.apache.org/guides/getting-started/index.html
     • Add new run configuration
       • Name: Jetty
       • Base directory: ${workspace_loc}/${project_name}
       • Goals: -X jetty:run
         • -X required for debugging, debugging works with this run configuration "Run", simply start it in debug mode
       • Details: http://docs.codehaus.org/display/JETTY/Maven+Jetty+Plugin
7. Run Package
   • The file /target/52n-uvs-1.0-SNAPSHOT.war is created, as are some other directories
8. Run Jetty
   • The service is available at http://localhost:8080/UVS/

Tasks

• Programming [HS]
  • move text snippets to a properties file in src/main/resources, e.g. en_EN.properties, see http://docs.oracle.com/javase/tutorial/i18n/resbundle/propfile.html
  • integrate i18n for JSP pages [DN]
  • add "no results" information for combinations with no results, no blank page for the user! [HS]
  • add properties file and sample.properties for configuration, for example the Google Analytics account ID, the server URL, ... then maven should automatically replace variables added as "${var.name}" in all files [DN]
  • Disable Google Analytics when developing (add a switch to some properties file?)

Tentative Ideas

• Feedback loop ("This suggestions was useful for me"), kind of an integrated survey -- DanielNuest - 2012-01-04
• Do detection of appropriate method solely on the original data which is uploaded (or size-limited subsets of the data) -- DanielNuest - 2011-12-05
  • requires the possibility of reading multiple file formats and detecting the attribute scale automatically
• Integrate feedback loop ("this was a useful suggestion", "I used this one and it worked" kind of user feedback) -- DanielNuest - 2011-11-17
• RESTful API mit JSON/XML responses
  • http://host.org/UVS/attribute/categorical/raster

Changelog

Version 0.x

• links to the scientific resources added to the visualisation description
• add small images (100x100 pixel) to the results, using tinybox: http://sandbox.scriptiny.com/tinybox/
• use tinybox also for the "more..." popup
• track usage statistics with Google Analytics

Developers

• Hansi Senaratne
• DanielNuest