CrossVis is an interactive visual analytics tool for exploring multivariate data.
CrossVis is written in Java and runs on Mac OS X, Windows, and Linux operating systems. The application is developed and maintained by the Oak Ridge National Laboratory Computational Data Analytics Group. The lead developer is Dr. Chad A. Steed.
CrossVis is an evolution of the EDEN visual analytics tool. If you are using CrossVis for your work, please use the following citation:
- Chad A. Steed, Daniel M. Ricciuto, Galen Shipman, Brian Smith, Peter E. Thornton, Dali Wang, and Dean N. Williams. Big Data Visual Analytics for Exploratory Earth System Simulation Analysis. Computers & Geosciences, 61:71-32, 2013. http://dx.doi.org/10.1016/j.cageo.2013.07.025
CrossVis has the following software dependencies:
- Java SDK or JRE version 1.8 or higher
- JavaFX version 11.0.2 or higher (https://openjfx.io/).
- Simple Logging Facade for Java (SLF4j) - a logging abstraction framework for Java (http://www.slf4j.org/).
- Apache Commons Math - a Java library of mathematical functions (http://commons.apache.org/proper/commons-math/).
- ControlsFX - a collection of high quality UI controls and other tools that complement the core JavaFX distribution.
Compiling the CrossVis Source Code
Compiling CrossVis is straightforward. The first step is to clone the repository. We supply a Maven POM file to deal with the dependencies. In the Eclipse development environment, import the code as a Maven project and Eclipse will build the class files. IntelliJ IDEA can import the code as a Maven project through similar procedures.
To compile CrossVis on the command line, issue the following commands (these commands assume that the Java SDK and Maven are properly installed on the development system):
$ mvn clean $ mvn compile $ mvn package
These commands should produce several build files in a new subfolder named 'target'. A jar file will be built (named 'crossvis-X.X.X-jar-with-dependencies.jar') with all dependencies bundled and the main class set. In Windows or Mac, this jar file can be double-clicked to execute the main CrossVis GUI. However, running CrossVis this way will use the default Java JVM memory allocations, which are usually too small for even moderate sized files. The preferred option would be to use the script files provided in the main project folder to run the application with a more suitable Java JVM memory allocation.
Due to issues with the state of the javapackager utility, the installer build will fail. However, the 'crossvis-X.X.X-jar-with-dependencies.jar' file should build successfully. It can be found in the target directory after the mvn package command.
First download the JavaFX SDK from https://gluonhq.com/products/javafx/ and unzip the file on the file system. Note the path to the unzipped folder for updating the launch script.
The compilation commands above will generate 2 jar files in the target directory. As noted above, an issue with the javapackage program will cause the installer build to fail, but CrossVis can still be executed using the included scripts. Copy the jar file with dependencies ('crossvis-X.X.X-jar-with-dependencies.jar') into the scripts directory.
Next, open either the crossvis.bat or crossvis.sh file and replace the placeholder '[[path_to_JavaFX_sdk]' with the path to the unziped JavaFX SDK folder. Be sure to leave the '/lib' at the end of this path. Then, run either the crossvis.bat script (Windows) or the crossvis.sh script (Mac or Linux).
The CrossVis window should appear after issuing this command. Example data files are provided in the data directory for getting started.