Analytical Visualization Framework a visual data processing and knowledge discovery system

1
Analytical Visualization Framework - a visual
data processing and knowledge discovery system

• Ivan Denisovich, Serge Shumilov
• Department of Computer Science III, University of
  Bonn

S. Shumilov Zürich
2
Motivation

• GLOWA Volta Project
• Integration of applications and services
• data transformation
• heterogeneous data
• various data sources
• Effective visual data exploration
• interactivity
• tracing of decisions
• specialized visualization methods
• Customized exploration interfaces
• different interfaces for different user groups
• easy programming and adaptation

• Solutions for analytical visualization
• GIS, statistical and spreadsheet software
• Programming libraries for visualization
• Visualization components frameworks
• Workflow composition systems

3
Analytical Visualization Framework

• Visual programming facilities
• data transformations
• visual interfaces
• visualization components
• Customized interactive data analysis tools and
  components
• Data exploration

4
Demonstration
5
Demonstration
6
Application M³ WATER model
geographic datahistoric and modeled climatic data
Hydrological model (MIKE Basin)
transformation
water supply
Visualization and Analysis (AVF)
water demand
transformation
Economical model (GAMS)
demographic and economic indexes water use
policies

• AVF is used for
• Access to distributed models
• Data collection and aggregation
• Customized presentation interface

7
Architecture

• Engine
• connection of modules
• notification about changes
• caching of intermediate results
• persistency of designs
• base classes API
• Visual programming workspace
• graph editor
• interactive manipulation
• monitoring and data exploration
• Library of modules
• generic computation
• data access
• spatial operations

AVF Module Library
JTS
Standard
Visual programming workspace
GeoTools
Groups
GRIA
User-defined
API
XML
XML
AVF Engine
Java Virtual Machine
8
Key features

• Interactive data manipulation
• changes immediately reflected
• Explicit transformation chains
• tracing of actions
• multiple paths

• Modular architecture
• input and output connectors
• Computation on demand
• change ? notification
• demand ? computation

1. Change
2. Notification
Module A
Module B




4. Computation
3. Request
Result
9
Benefits and Drawbacks

• Generic reusable modules to build complex
  transformations
• Visual programming of data transformations and
  interfaces
• Full history of applied transformations. Changes
  in parameters of intermediate steps result in
  direct recomputation
• High code portability ensured by Java

• Designs can grow large and tangled
• Predefined programming paradigm (execution,
  communication, interfaces)
• Inappropriate if several parameters should be
  changed for a new result
• Third-party non-Java visual libraries are hard to
  integrate into the framework

10
Thank you for your attention!
Questions?
divan_at_iai.uni-bonn.de www.glowa-volta.de