InfVis Applet Manual


(Version 2.4.10)
Frank Oellien


Table of Contents


1. Introduction
2. Applet Parameters
2.1 The datafile Parameter
2.2 The displayStyle Parameter
2.3 The dataBackground Parameter
2.4 The coordStyle Parameter
2.5 The showZeroPlane Parameter
2.6 The Dimension Parameters
2.7 The Dimension Style Parameters
3. The GUI Components
3.1 The Menu Bar
3.1.1 The View Menu
3.1.2 The Data Menu
3.1.3 The Options Menu
3.1.4 The Help Menu
3.2 The Data Window
3.3 The Control Panel
3.3.1 The Mapper Panel
3.3.1.1 The General Settings Mapper Panel
3.3.1.2 The Axis Mapper
3.3.1.3 The Size Mapper
3.3.1.4 The Shape Mapper
3.3.1.5 The Color Mapper
3.4 The Tool Panel
3.4.1 The Navigation Tool
3.4.2 The Dynamic Filter Tool
3.4.3 The Selection Tool
3.4.4 The Detail Tool


1. Introduction

Today, chemical companies are routinely employing synthesis technologies such as High-Throughput-Screening, Combinatorial Synthesis and Parallel Synthesis that generate terabytes of data per year. These datasets are highly relevant for chemists, because they potentially deliver insights into chemical and biochemical trends and principles and could lead to faster development and higher numbers of drug candidates. However, the development of data mining and visualization tools are capable of analyzing these large amounts of multi-dimensional data appears not to have been able to keep pace with the dramatic increase in size of these datasets. Ultimately, this situation has become one of the most critical bottlenecks in chemical R&D today. This applet allows an interactive and graphical data mining and visualization of large-scale multi-dimensional datasets. Beside the applet version it also exists an application version with local file and database connection support to import own datasets easily.

2. Applet Parameters

The applet provides several applet parameters that allow some initial settings.

2.1 The datafile Parameter

<PARAM name="datafile"value="./path/to/your/file.txt">

The datafile parameter is essential because it includes the path to the dataset file. If this parameter does not exist an error message will result.

2.2 The displayStyle Parameter

<PARAM name="displayStyle" value="barchart">

The displayStyle parameter will set the initial display style (see 3.1.1). The possible values are ‘barchart’, ‘scatter’ and ‘glyph’.

2.3 The dataBackground Parameter

<PARAM name="dataBackground" value="#000000">

The dataBackground parameter allows to set the background color of the Java3D window. The color string (value) can be written by using name of the color or by using the hexadecimal HTML representation.

2.4 The coordStyle Parameter

<PARAM name="coordStyle" value="open">

The coordStyle parameter allows the user to decide if the coordinate system box will be displayed in a closed or open style. The possible values are ‘open’ and ‘closed’.

2.5 The showZeroPlane Parameter

<PARAM name="showZeroPlane" value="yes">

By using the showZeroPlane parameter the user can choose if an additional Y zero plane is visible in the coordinate system. The plane will only be visible under right conditions (see 3.1.3).

2.6 The Dimension Parameters

<PARAM name="xDimension" value="0">

<PARAM name="yDimension" value="1">

<PARAM name="zDimension" value="2">

<PARAM name="sizeDimension" value="-1">

<PARAM name="colorDimension" value="2">

<PARAM name="shapeDimension" value="-1">

The dimension parameters can be used to map dimensions to the axis or object mappers individually. The values have to be integers that correspond to the column numbers in the data file (starting from 0). When a mapper should not contain a dimension the -1 value can be chosen. Usually these parameters are not necessary because the dimensions will be mapped automatically. However, they can be useful sometime, e.g. to map no dimension to a mapper or to chose another order than the column order in the data file (see 3.3.1).

2.7 The Dimension Style Parameters

<PARAM name="xDimStyle" value="categorical">

<PARAM name="yDimStyle" value="numerical">

<PARAM name="zDimStyle" value="categorical">

<PARAM name="sizeDimStyle" value="numerical">

<PARAM name="colorDimStyle" value="categorical">

<PARAM name="shapeDimStyle" value="categorical">

The dimension style parameters are used to define if a data dimension should be displayed in a categorical or numerical style (Only numerical data can be displayed in both styles!). The values are ‘categorical’ and ‘numerical’ (see 3.3.1).

3. The GUI Components

3.1 The Menu Bar

The menu bar contains in the current version four submenus: View, Data, Options and Help.

3.1.1 The View Menu

The ‘View’ menu contains the most important view options like choosing the display style, selecting the visible panel or choosing the 3D viewpoint.

The ‘Display Styles’ submenu allows the selection of the display style. Currently, the user can choose between glyph, 3D bar-chart and scatter representation. The 3D scene will be updated immediately.

     

The ‘Maximize Render Panel’, ‘Control Panel’ and ‘Tool Panel’ option are used to hide or show the control and tool panel. By using the ‘Maximize Render Panel’ option the render panel will be displayed using the whole application window.

The ‘3D Viewpoints’ submenu allows the user to choose between different data viewpoints – Standard View, Front View, Side View and Bird View (see 3.4.1).

3.1.2 The Data Menu

The ‘Data’ menu allows fast access to all data related functions like dataset information and settings and for import of datasets.

By using the ‘Dataset Info/Settings’ option the user can open a dialog window that allow the modification of dataset settings like dataset name, selection of the currently displayed dataset, deletion of datasets.

3.1.3 The Options Menu

The ‘Options’ menu contains only one menu item in the moment – the ‘General’ option. Clicking on this item will open the options dialog window.

The options dialog contains several buttons to control color of background, color of coordinates (axes and labels), color of the detail box, color of the zero plane and the selection color. It also contains a switch option to choose if the coordinate system will displayed in an open or closed style.

   

Finally, the user can choose if he wants to add an additional Y zero plane to the 3D scene. This option is only available when numeric data is mapped to the y axis. It is useful when the dataset contains negative and positive values.

3.1.4 The Help Menu

Like the ‘Options’ menu the ‘Help’ menu contains only menu item – the ‘About ’ button. The corresponding window contains version number information.


3.2 The Data Window

The data window consists of the Java3D render panel and a range slider for each orthogonal axis. The names of the dimension that have been mapped to the axis will be displayed on the range sliders. Furthermore a tool tip will appear when the mouse pointer is moved over the range slider. The range sliders can be used to hide parts of the datasets. Because Java3D take advantage of 3D graphic hardware, the interactivity and number of data points depends strongly from the used 3D graphic card. The user can navigate through the scene by using the mouse:

Left mouse button + mouse movement: rotation
Left mouse button + ALT + movement: zoom
Right mouse button + mouse movement: translation
Left mouse button + Shift (on 3D object): Detail
Right mouse button + Shift (on 3D object): detail overview
Left mouse button + STRG (on 3D object): selection
Right mouse button + STRG (on 3D object): selection overview

3.3 The Control Panel

The control panel consists of two subpanels – the mapper panel and the status/error window.

3.3.1 The Mapper Panel

The mapper panel handles the mapping of the data dimension to the three orthogonal axes, the size, color and shape of the 3D objects. After loading a dataset the application will automatically map the first six dimensions to this six ‘mappers’. Of course the user can modify these settings again. It is also possible to map no dimension to a mapper. In this case standard values will be used. Because of the limited number of mappers, only six data dimension can be mapped to them. However, additional data dimensions (>6) can also be included by using the dynamic query devices (see 3.4.2). The mapper panel contains a general settings panel and also a specific settings panel for the axes, the size, the color and the shape. The user has to use the ‘Update 3D Scene’ button at the bottom of the mapper panel if the changes in the settings panels should be represented in the 3D scene.

3.3.1.1 The General Settings Mapper Panel

The ‘General Settings’ panel gives an overview which data dimensions are currently mapped to the mappers. By using the pull-down lists the user can select every data dimension (except meta information dimensions) that exists in the dataset and map it to the selected mapper.


3.3.1.2 The Axis Mapper

Like the size, color and shape mapper the axis mapper includes a dimension info section and several settings options. The ‘Dimension Info’ section shows the currently mapped data dimension, the number of different values in this dimension, the minimum and maximum values and the data type (numeric or non-numeric). Within this section the user also can change the mapped dimension. In the case of numeric data the minimum and maximum fields contain the numeric minimum and maximum. Otherwise the minimum is 1 and the maximum represents the number of single data points.

Beside the ‘Dimension Info’ section there is a ‘Value Settings’ section available. This section shows how the data is handled. In the case of non-numeric data only a categorical representation is possible, the range type is tight and all fields can not be modified. However, when the data is numeric the user can choose between a numerical or a categorical representation. By choosing the numerical representation the user can also modify the other fields and options. The data range can be loose or tight. The tight representation will fit the axis to the minimum and maximum of the data, the loose representation will choose a nice number range for the axis (e.g. 0 – 100 instead of 2 - 93). In this representation the user can also modify the minimum and maximum values. The ‘Labels & Tick Marks’ section allows to choose the distance between the ticks and if labels should be painted or not. The ‘Axis Length & Axis Scale’ section is currently limited to linear scales. Future releases of this application will also allow a logarithmic scale. The axis slider allows the user the modification of visible axis length.

3.3.1.3 The Size Mapper

The size mapper also includes the same dimension info section like the axis settings panel. In the case of numeric data the user can also choose between a numerical and a categorical representation, however, the numerical representation is recommended. In the numerical representation the numeric values will be used to calculate the corresponding size otherwise the number of different sizes corresponds to the number of single values. By using the minimum and maximum sliders the user can define the visible minimum and maximum sizes of the objects. The sizes will than be computed beside these limits. In the scatter display style the size settings panel is not available.

3.3.1.4 The Shape Mapper

The shape mapper only includes the dimension info section in the moment. Settings options will be implemented in future releases. Only a categorical representation is possible. In the moment we provide the following shapes: a) glyph style: sphere, box, cone, cylinder (max. 4 different data values); b) bar-chart style: cylinder and box (max. 2 different data values); c) scatter style: The shape settings panel is not available, because only points are represented. If a data dimensions contains more than 4 or 2 different values it can not mapped to the shape mapper. In such cases the shape mapper uses standard values (sphere or cylinder).

3.3.1.5 The Color Mapper

The color mapper controls the color related options. Non-numerical data dimensions can be represented if they contain maximal 12 different single values. This limit has been chosen, because nominal information coding should allow easy recognizing. The application uses the 12 recommended color for coding: red, green, yellow, blue, black, white, pink, cyan, gray, orange, brown and purple. However, the user can also define every RGB color for coding. By clicking on one of the categorical color buttons, the categorical color chooser window will appear. The user can chose for each data point a specific color.

If no data dimension is mapped to the color mapper, the data points will be represented in a standard color, that can be modified by the user. Numerical data can also be represented in a categorical way (up to the maximum of 12 different values). However, it is more reasonable to use the numerical representation in this case.

The user can choose between different three standard color gradients and two definable color gradients to represent the data.

Standard color gradients:

Spectrum Color Gradient

Rainbow Color Gradient

Primary Color Gradient

Definable color gradients:

Two Color Gradient

Three Color Gradient

All color gradients allow the definition of maximum and minimum values. These limits can also be used to define value thresholds. In contrast to the standard color gradients the definable color gradients also allow the definition of minimum and maximum colors.

Furthermore the three color gradient also allows the definition of a “zero” color and a “zero” value. This value can be defined everywhere within the data range.

3.4 The Tool Panel

The tool panel is the most important panel of the application, because it contains all the tools which provide the interactive and dynamic datamining of the datasets. The user can chose between navigation tools, dynamic filter tools, selection tools and detail tools.

3.4.1 The Navigation Tool

This tool panel contains the same standard 3D viewpoints like they have already described in the view menu (see 3.1.1).

 

Standard Viewpoint

Front Viewpoint

 

Side Viewpoint

Bird Viewpoint 

3.4.2 The Dynamic Filter Tool

The filter panel allows the creation of data dimension filter. This option is not limited to six data dimensions like the mapper. A filter can be defined for every data dimension that is available in the dataset. This approach allows the analysis of multi-dimensional datasets with more than six dimensions. The user can choose between item sliders, range sliders, check boxes and radio button based filters. Only one filter can be defined per data dimension. The filter can be added by using the ‘Add Filter’ button. The filters will be added in the ‘Filter List’ panel. Furthermore it is easily possible to remove specific filters by using the ‘Remove Filter’ button.

All changes on the filter tools will affect immediately the corresponding 3D scene. Therefore this method allows a fast and interactive refinement of the dataset.

3.4.3 The Selection Tool

The selection tool provides two methods for data point selection: a) selection by using selection boxes and b) manual selection of data points with the mouse. Both methods can be combined to allow a comfortable subset selection. Filter tool settings (see 3.4.2) will also be taken into account.

Selection Boxes:

Selection by selection boxes is very comfortable, because the user can define several moveable semitransparent selection boxes which select all data points in there space (‘Add Selection Box’ button). The frontiers of the boxes can easily be defined by using range sliders.

Manual Selection:

In contrast to the selection box method the manual selection can not be added through a button. However, it is not more difficult. The user only has to press the STRG key and the left mouse button pointing on the desired 3D object in the render. Around the object a selection box will appear and the selected object shows the selection color (e.g. yellow). Within the tool panel a selection list will appear that contains all selected objects. All selected objects are represented by their data point index and the mapped dimension values. Within the list the user can click on a selected data point. The corresponding 3D object will be highlighted by the selection box. This allows an easy identification of single data points in the 3D scene. Of course the user can also delete marked data points from the list by using the DEL key. Generally more than one data point exists on a specific position in the coordinate system, but it is not visible because a other data point object has a bigger size. To get access to the underlying data points the user has to press STRG and the right mouse button. Now a ‘Specify Selection’ window will appear. It contains all data points that lie in the users view direction.

The user can move through the list. The currently marked data point is highlighted in the scene by the selection box and in the tool panel the detail tool (see next chapter) shows all information about the marked data point. After the user has found the desired data points he can use the ‘Add Selected Datapoints’ button to add the data points to the selection list. The selection box selectors and the manual selector can be easily deleted by using the ‘Remove Selector’ button.

3.4.4 The Detail Tool

Another very important tool is the detail tool. The detail tool gives the user access to all data values for a specific data point. By pressing the Shift key and the left mouse button during pointing on a 3D object, the user can open the detail window. The picked object will be highlighted with the selection box. The corresponding data point information will be shown in the detail window.

 

As already described in the last chapter, it can sometimes be difficult to get information about all data points, because they lie within other data points. By using the Shift key and the right mouse button a ‘Specify Detail’ window will appear that gives the user also access to the hidden data points.

 

As already mentioned earlier the application is capable to handle meta information. Meta information can be used to embed information that will not be displayed in the render window like encoded images or URLs. This information will be displayed within the detail panel. Currently the application can read Base64 encoded images and URLs. The images will simply be painted within the detail panel. Embedded URLs are directly accessible through buttons. By pressing this buttons the OS standard browser will be launched with the corresponding URL as target.