Intuitive information visualization and direct data editing

ILOG Discovery is a visualization tool for analyzing data sets intuitively and communicating findings within an enterprise information system. It also serves as an interactive tool for browsing and editing databases, using a direct manipulation paradigm instead of a form-based query system. It is available for free for a limited time period, while some of its features are made available for professional uses in the JViews charts module.

Complete analysis tool

Powerful preset visualizations let you start analyzing data sets easily, a general editing scheme lets you define views that best highlight your findings, and specialized navigation and selection features let you find trends, pinpoint exceptions and reveal the structures that govern your data.

Data visualization component

A Java component using the tool's features can feed visualizations with data from enterprise information systems and share them on enterprise portals, or let you integrate ILOG Discovery features in your application.

Discover the next level in visual data analysis

ILOG Discovery applies a new dataflow model that provides access to a very large class of visualization algorithms through interactive parameter settings. The dataflow model has a fixed number of parameters for ease of use and predictability, and co-routine-like mechanisms for declaring visualization features that require state handling.

On this web site, you can:download the free preview software, find additional material on the technology and its uses; walkthrough some typical usage sessions, explore typical uses of ILOG Discovery for visual data analysis or direct data editing and edition ; and provide comments and wishes. ILOG Discovery is a concept software whose goal is to proof new technologies before their introduction in the ILOG JViews product line.

Quick readers, can glance at its main features:

Purpose: Information visualization tool ; visual data analysis and reporting of business data ; direct Direct data editing and editing.

Data Handling

Data tables such as spreadsheet tables, SQL tables, web log files... custom readers can be integrated to load any sort of data table. Can handle data sets from 10000 to 1 Million lines and about 30 columns, depending on the hardware platform and type of view.

Load data tables from various ASCII text formats, such as spreadsheets, log files...
Direct connection to data storage systems, via data source API.
Retouch data sets locally: add computed columns, perform local changes, create sub tables via a logical expression or a manual selection.
Load and manipulate multiple data sets.
Reload feature to reopen data sets that may have been modified.
Export data tables.
Accepted data types: String, Enumerated, Integer, Number, Date and any serializable java class.

Visual Data Analysis

Powerful, ready-to-use visualizations:
- Overviews (histograms and distributions) to get a sense of content.
- Graphs to analyze trends and find exceptions, to refine your sense of content,
- Data Maps and Hierarchies to categorize, hierarchize and distribute, get a sense of structure.
Fully customizable visualizations, to define a much broader range of visualization types, in a unified way. Most notably allows crossing visualization features (such as grids of 2D plots).
Navigation: Geometrical: Zoom, pan ; Semantic: focus in/out ; view exact values with object inspector, table viewer and tooltips.
Selection and marks to perform and display cross-view selection.
Parallel analysis: Multiple data sets loaded and displayed at the same time. Multiple views visible at the same time.
Easy window manipulation: views are automatically iconified to make room for new views. A view shelf provides direct access to an unlimited number of views. Non-overlapping window schemes avoid all window manipulations.
Macros mechanism to perform combined settings in one single action.

Extensive Visualization parameters

A patented algorithm enables combining the following parameters and define recursively sub visualizations. Those parameters enable creating through a simple point and click interface all data-linear visualization algorithms.

Partitioning: Partition and cluster on enumerated, numeric, date values, URL and IP addresses. Fully recursive model.
Sort
Layout: Rectangle, Ellipses and Polylines shapes
- Automatically adjusted Scales
Frame and margins
Labels
Decoration: base decoration with support for qualitative, quantitative and diverging color scales. HSB color model, gradient color model, transparency attribute.
Per record visualization
- Field display: per line bar charts and "bar code".
- Interactive filters enable dynamic queries.
Sub visualizations enable fine tuning views to highlight relevant findings while maintaining the context.

Project Management

Project management features let you store and retrieve data set references, views (data set reference + visualization), visualizations, marks (selections of lines in a data set), notes...

Report and share results

Export and share results: Export views and data sets to other tools: presentation, spreadsheet… Deploy views and full projects on enterprise portals, with applets instantiating the ILOG Discovery components.

Up to date user interface

Full undo/redo, Integrated online help, Preferences handling, Fully internationalized

Component

A Java component tightly coupled to the interactive tool lets customize those views, connect them with up-to-date data from enterprise information systems and deploy them on enterprise portals.

News

15/05/07: Two videos showing Discovery main features are released: the visualization model, and direct data editing.

18/10/06: paper on Direct Data Editing presented at UIST06.

15/04/06:ILOG JViews Charts now delivers some of the best features of ILOG Discovery: treemaps and small multiple views, alongside its wide range of classic charts. You may want to check out or evaluate this new treemaps and charts java component gui library.

15/04/05: new research paper on direct data manipulation with ILOG Discovery.

15/03/05: Version 1.07 includes:
- removes parallel histograms
- adds other types of overviews
- graphic annotations

15/02/05: Version 1.06 includes:
- instant project deployment
- extended selection tools
- more data editing functions

27/11/04: Version 1.05 includes:
- view creation wizard
- squarified treemaps
- editable tables...

9/10/04: see ILOG Discovery at the Information Visualization Software Infrastructures workshop at IEEE 2004 Visualization conference.

28/09/04: API available to selected users.

27/07/04: Mac OS X version

25/06/04: using ILOG Discovery for source code quality management at JavaOne 2004.

14/05/04: First plugin: ILOG Visual CP

28/04/04: see ILOG Discovery at the CHI 2004 conference.

1/03/04: first usage survey.

1/12/03: Discovery preview released.