WO2001093204A1 - Generating image data - Google Patents

Abstract

Data items (201) having at least four numerical fields (211, 212, 213, 214) are processed. A position for each data item is defined by processing the first numerical field and the second numerical field to generate a first position co-ordinate and a second position co-ordinate. A visual characteristic for each of the data items is defined by combining weighted sums of the contributing numerical fields. An image is rendered so as to represent the data items as a three-dimensional view by combining the first co-ordinate and the second co-ordinate with the visual characteristic.

Description

Generating Image Data

Field of the Invention

The present invention relates to generating image data from data items wherein said data items have a plurality of numeric fields. In particular the invention relates to generating image data by processing numerical data having a plurality of dimensions.

Introduction to the Invention An abundance of numerical data is available for analysis using a wide variety of mathematical methods. The abundance of data has arisen from the implementation of systems capable of digitally sampling real world events, and of storing an unlimited amount of this data. Many systems yield useful information as a result of the mathematical analysis that may be performed upon data of this kind. For example, several types of statistical analysis may be performed on numerical data created when measuring performance of critical machinery such as jet engines, new materials for constructing buildings, or the modelling of complex systems such as telecommunications networks. Historically, computers have been successfully used in numerous scientific and mathematical applications. However, certain natural systems have been found to be largely resistant to numerical analysis. The study of weather systems and financial markets has led to the development of chaos theory and the study of the behaviour of unpredictable systems. A related field is the study of fractals, which aims to predict the behaviour of chaotic systems using mathematical tools outside those of pure statistics. Although progress in chaos theory and fractals has improved the predictability of complex systems behaviour, these new mathematical tools have also revealed fundamental limitations in the degree to which the behaviour of such systems can be predicted.

Although complex systems analysis has benefited immensely from the introduction of computers, a significant human input is still found to be necessary. In financial markets, an experience of stock market fluctuations over many years enables an individual trader to make quick intuitive decisions that are considerably better judged than any formal analysis could provide. The ability to perform in this way requires considerable awareness of many aspects of financial markets, as well as experience and familiarity with mathematical procedures, in order to anticipate trends.

An example of this kind of trading is arbitrage. In arbitrage, a trader may buy a first currency, for example dollars, convert the dollars to a second currency, for example yen, and then convert the yen back into the original currency, such as Euros. Depending upon market conditions, it is possible to finish with more Euros than the trader originally started with. A trader experienced in arbitrage can spot combinations of trends, behaviours and conditions, and quickly make a sequence of trades that results in a profit. This type of action requires an acute awareness of market conditions as they change in real time. While computers are not capable of generally predicting the behaviour of complex systems, in financial trading they have a central role to play in the presentation of information. With a large high definition monitor and a fast network connection, todays analysts and traders have instant access to whatever financial information they require. Due to the diversity of data, information of this type is usually supplied in the form of numbers. A particular sequence of numbers, for example share prices in a particular company or stock exchange index, can be converted into a graph using readily available spreadsheet graphing facilities. Complex numerical analysis tools may be applied to such data, including the fractal analysis methods developed by Benoit Mandelbrot and others. The results of analysis can also be graphed without difficulty.

These methods are widely applied in stock markets. However, due to the complexity of data that is available, there is no known way of providing a unified view of diverse data, beyond that of the simple graph. This is a problem inherent in the data itself, resulting from its origination in a complex dynamic system.

Summary of the Invention

According to an aspect of the present invention, there is provided a method of generating image data from a plurality of data items, in a data processing system including memory means for storing instructions and said data items, and processing means for generating said image data; wherein said instructions define operations to be performed by said processing means to generate said image data from said data items; said data items having at least four numerical fields, including first and second numerical fields and a plurality of contributing numerical fields; and said operations include repeated steps of: defining a position for each of said data items by processing said first and second numerical fields to generate first and second position co-ordinates; defining a visual characteristic for each of said data items by combining weighted sums of said contributing numerical fields; and rendering image data representing said data items in a three dimensional view by combining said first and second co-ordinates with said visual characteristic.

Brief Description of the Drawings

Figure 1 shows the Internet, including a data service provider supplying data items to users at several connected computer terminals;

Figure 2 summarise the present invention, in which image data is generated from data received from the data service provider shown in Figure

1; Figure 3 details a terminal of the type shown in Figure 1, including a computer and a monitor;

Figure 4 details hardware components of the computer shown in Figure 3, including a main memory;

Figure 5 summarises steps performed at the computer terminal shown in Figure 2, including a step of generating image data;

Figure 6 details contents of the main memory shown in Figure 4 during the step of generating image data shown in Figure 5;

Figure 7 concurrent steps performed during the step of generating image data shown in Figure 5, including a step of generating a view; Figure 8 summarises the views that may be generated during the step of generating a view shown in Figure 7, including a grid view, a planets view, a nails view, a particles view, an a top list view;

Figure 9 details data of the type received from the data service provider shown in Figure 1; Figure 10 details a data item resulting from combining data of the type shown in Figure 9;

Figure 11 details a first aspect of the grid view indicated in Figure 8, including a first numeric field and a second numeric field;

Figure 12 details a second aspect of the grid view shown in Figure 8, including weights for additional numeric fields;

Figure 13 details the planets view indicated in Figure 8;

Figure 14 details the nails view indicated in Figure 8;

Figure 15 details the particles view indicated in Figure 8, at an initial condition;

Figure 16 details the particles view indicated in Figure 8, at a final condition;

Figure 17 details the top list view indicated in Figure 8; Figure 18 details the process for generating a grid view as identified in

Figure 11;

Figure 19 summarises steps required to generate the grid view shown in Figures 11 and 12, including steps of sorting a grid and rendering a grid;

Figure 20 details the step of sorting a grid shown in Figure 19, including a step of comparing data items;

Figure 21 details the step of comparing data items shown in Figure 20;

Figure 22 details the step of rendering a grid shown in Figure 19, including calculating a Y value; and

Figure 23 details the step of calculating a Y value shown in Figure 22.

Detailed Description of The Preferred Embodiment

The invention will now be described by way of example only with reference to the accompanying drawings.

Figure 1

An environment for distributing data is shown in Figure 1. A data service provider 101 supplies data in response to requests for data received from users on the Internet 102. Users are connected to the internet via Internet service providers (ISP) 111 to 116. Additionally, certain users may connect to an Internet service provider 114, 115 via an intranet

116 or 117, that is provided by a company or organisation for internal data communications. Users have computer terminals 121 to 135 that receive data from the data service provider 101. Data from the data service provider 101 is supplied to computer terminals where the data is stored temporarily in the form of data items, each of which has a plurality of numerical data fields. In the present embodiment, each data item represents financial information about a company or other financial entity.

Data items comprise a plurality of numerical data fields that define various characteristics of a company's financial condition. These numerical data fields are collectively known as a company's fundamental data, or fundamentals. The fundamentals used by financial analysts or stock market traders in order to make informed decisions about market purchases and sales.

The data service provider 101 supplies data in a data file standard known as comma separated values (CSV). A CSV file, once downloaded from the internet 102, can be imported into a spreadsheet application operating on any of the computer terminals 121 to 135. Thereafter, a user may configure the spreadsheet to perform known numerical analysis methods, and use built-in data visualisation functions to display graphs. However, the complexity of company fundamentals, comprising in the region of fifty or more parameters for each individual company, ensures that presenting the information is not, in itself, sufficient to enable a user to gain insight into hidden aspects of market behaviour. Professional traders usually rely on an intimate knowledge of market behaviour gained over many years of experience in order to distinguish subtle trends and characteristics that may be of importance.

Figure 2

The invention filters out an unnecessary level of complexity inherent in the numerical data of a complex dynamic system such as a stock market. The invention is summarised in Figure 2. Incoming data 201 is received from the data service provider 101. The incoming data 201 includes data that includes company fundamentals. The user is interested in a subset of companies for which data has been received, and these are denoted company A, company B, company C and so on. Each company has around fifty fundamentals, of which the user is only interested in a small number. These are considered as numerical fields 211, 212, 213, 214 and 215. These numerical fields may be any of the company fundamentals, and, in some cases it may be advantageous to use the same company fundamental in more than one of these numerical fields.

The first and second numerical fields 211 and 212 are processed to define positional co-ordinates in a two dimensional plane. A two- dimensional array 231 stores company identifiers, or pointers. A sorting process is used that results in ascending values of the first numerical field 211 being arranged from left to right in the array, and ascending values of the second numerical field 212 being arranged from bottom to top in the array. Co-ordinates in the array are then used to define data item positions in a two-dimensional XZ plane.

The third, fourth and fifth numerical fields 213, 214 and 215 are combined by a weighting process to define a visual characteristic. In the preferred embodiment the visual characteristic is a height, Y, of a column 241 representing a data item. X, Y and Z values are combined by rendering in a three-dimensional view. If the third, fourth and fifth numerical fields are Q, R and S, and their respective weights are A, B and C, then the equation used to combine them is:

Y = A.Q + B.R + C.S Figure 3

A computer terminal of the type shown at 121, operated by the user while receiving data from the data service provider 101 , is illustrated in Figure 3. A computer 301 contains processing and other circuitry required for connection to the Internet. The computer 301 receives commands from the user in the form of keystrokes made on a keyboard 302, and in the form of movements and button presses made on a computer mouse 303. Various forms of processing are performed by the computer 301 in order to interpret these commands, and this interpretation is dependent upon the context or state of the computer at the time the command is given. In this way, the input devices 302 and 303 provide a general means of user command input, that may be applied also to an embodiment of the present invention. In response to user commands, and the internal state of the computer 301 , image data is converted into signals supplied to a monitor 304, resulting in a display of images responsive to user operations and computer states.

Figure 4

The computer 301 shown in Figure 3 is detailed in Figure 4. A central processing unit (CPU) 401 performs processing in response to instructions stored in a main memory 402. The main memory comprises sixty-four megabytes of dynamic RAM. The central processing unit includes primary data and program cache circuitry to minimise the transfer time of repeatedly accessed data and instructions. A suitable processor is a Pentium II! available from Intel Corporation. Secondary cache circuitry, also provided in the CPU 401, operates at a lower speed, but at a larger memory capacity, thereby minimising transfer of data over a bus 403 electrically shared by several other components of the computer. A hard disk drive 403 provides non-volatile storage for CPU instructions and data, and further extends the effective capacity of main memory through the use of virtual addressing techniques. A graphics card 404 receives instructions for the rendering of image data generated by the CPU 401. In the graphics card 404, digital to analogue converters for each of red, green and blue colour components generate electrical signals that are supplied to the monitor 304 via an electrical connection 405. The graphics card 404 generates frames of image data at a resolution of 1024 by 768 pixels, where each pixel has a data depth of four bits for each of the red, green and blue color components. In the preferred embodiment, the graphics card 404 includes a processor configured to render image data supplied to the graphics card in the form of OPENGL graphics commands. However, in alternative embodiments it is possible for the CPU 401 to perform the processing required in order to implement OPENGL functionality, but at a slower speed.

OPENGL commands facilitate definition of shapes, objects and lighting in a virtual 3D space, that may be viewed in a conventional two dimensional window displayed on the monitor 304. A three dimensional window responsive to OPENGL commands is considered as having an OPENGL context, that in combination with other types of windows and contexts, facilitates user interaction with the virtual 3D world it contains.

Instructions for operations performed by the CPU may be installed from a CDROM drive 406. Instructions are usually stored in compressed form on a CDROM, and installed by a smaller number of installation instructions controlling the CPU 401 , so that new instructions are stored onto the hard disk drive 403. Thereafter, a user may initiate the new instructions by an appropriate context-dependent command on the keyboard 302 or mouse 303. CDROMs may also be used as a source of data, for example historical data relating to company performance. This historical data may extend to several hundred megabytes in size, if several years and several stock markets are represented. An alternative source of new instructions and or data is a local area network, such as a corporate intranet 116. Connectivity to an intranet is made via a network card 407, and an Ethernet connection 408.

Input output circuitry (I/O) 409 facilitates communication between the keyboard 302, mouse 303 and the CPU 401. A modem 410 facilitates connectivity to ISDN or a telephone telephone line. Internet access may be achieved either via the modem 410 or the network card 407.

Figure 5

Operations performed by the user of the computer terminal 121 are summarised in Figure 5. At step 501 the computer is switched on and the processing system, comprising the CPU 401 and the main memory 402, are initialised. At step 592, if necessary, instructions for generating image data are loaded from a local area network or CDROM 503. At step 504 the user initiates instructions for generating image data. Thereafter, image data is generated in response to contexts and states generated by user commands and incoming data 505 received over the Internet 102, from the data service provider 101.

The service provider 101 supplies data to the user terminal 121 at a rate dependent upon a previously agreed contractual arrangement. Depending on the importance of the data, updates may be received once a day, or virtually continuously, with updates being received every minute or so. Having completed generating image data, several context settings, defined by the user, are saved at step 506. These may then be recalled during a later session. At step 507 the processing system is shut down, requiring a synchronisation of main memory 402 and the hard disk drive 403, in order to ensure that all permanent data have been stored in nonvolatile media.

Figure 6

Contents of the main memory 402 while generating image data 504 are detailed in Figure 6. A Windows NT (TM) operating system 601 provides common instructions, such as those required for accessing the hard disk drive 403, graphics card 404 and all components of the computer. The operating system 601 includes instructions for facilitating OPENGL rendering commands in co-operation with processes performed by the graphics card 404. Instructions for generating image data 602 facilitate reception of data from the data service provider 101 , interaction with user commands supplied from the keyboard 302 and mouse 303, and generation of graphical contexts within which windows are generated for display on the monitor 304. These contexts include an OPENGL window, within which a projection of three-dimensional world space is viewed and adjusted. Other applications 603 include instructions for file management and data backup. Instructions 601, 602 and 603 are permanently stored on the hard disk 403, and are loaded on demand into the main memory 402. The operating system 601 is loaded automatically when the computer is switched on. Thereafter, sequences of instructions are selectively identified by the operating system for execution in short bursts of time, or time slices. This provides the illusion to the user, that the CPU 401 is executing multiple tasks concurrently. Furthermore, multi-tasking enables the computer to receive data updates 505 over the internet while at the same time generate image data 504 on the monitor 304 without the user being aware of any interruption.

During execution of instructions 601 to 603, portions of memory 402 are allocated and de-allocated as a result of changing instruction requirements. A dynamic workspace 604 is provided by the operating system 601 , in the form of a stack, to facilitate the allocation of memory in response to requests from applications 602, 603 as their instructions are executed. Other regions of memory are also allocated for specific requirements of the instructions for generating image data 602. An incoming data buffer 605 temporarily stores incoming data from the data service provider 101, received over the Internet. Data items 606 include data from the buffer 605 that have been formatted in such a way as to be suitable for generating image data. User preferences 607 include settings relating to generating image data, including the way in which data items 606 are generated from incoming buffered data 605.

Figure 7

The step of generating image data 504 shown in Figure 5 is detailed in Figure 7. At step 701 the user selects a view. At step 702 the appropriate view is generated, resulting in the generation of a three-dimensional view in a window on the monitor 304. Incoming data is received at step 703, and a database of updated stocks and shares of companies and other financial entities is stored in the incoming data buffer 605 in main memory. The user defines a data requirement at step 704, resulting in data items selected from the database 605 being stored as data items 606 in main memory 402. Selected data items usually include a subset of a much larger number of companies and financial entities whose fundamentals are stored in the buffer 605. The selected data items 606 are supplied as additional input data to the step of generating a view 702. Steps 701 to 704 each represent processing performed in a multitasking processing environment, as facilitated by the operating system instructions 601. These steps are not necessarily performed in any particular order, and at any time, the precise functionality implemented by any of these steps is defined by the data available to it. As an additional parallel process, step 705 saves user preferences to hard disk. User preferences include configuration data, that includes, for example, an indication as to which companies are to be selected as data items 606 when generating a view. Other data relating to specific views is also stored in association with a particular configuration, as will be detailed subsequently.

Figure 8

The step of generating a view 702 shown in Figure 7 is detailed in Figure 8. At step 800 a selection is made on the basis of the view identified by the user at step 701. Thereafter the appropriate view is selected from one of the following: grid view 801 , planets view 802, nails view 803, particles view 804, top list view 805 and other view 806. Flow lines from decision box 800 to process boxes 801 to 806 have been omitted for the sake of clarity.

Figure 9

An example of the data received from the data service provider 101 is shown in Figure 9. Figure 9 details contents of a file that has been supplied in the comma separated value (CSV) format. A first column 901 defines the full names of companies, or other financial entities, such as trust funds and so on. A second column 902 specifies a unique symbol for the company or financial entity, comprising a few alphabetic characters. A third column 903 specifies the share price for the company. Share prices may fluctuate rapidly, and so it is common to receive a CSV file in which the only company fundamental defined is the share price. Other CSV files may include other company fundamentals, updated less frequently. Several data service providers supply data of this kind, varying in the cost and frequency of updates. A suitable source of such data is quote.com, and information relating to CSV formats and other implementation details is available at http://www.quote.com/quotecom/qcharts/qfeed.asp?=qfeed.

Figure 10

A data item comprises many company fundamentals, assembled from information supplied, perhaps with frequent updates from several CSV files. A resulting data item is shown in Figure 10. A first column 1001 shows the name of the fundamental, and a second column 1002 shows the value of the fundamental. Typically, many thousands of companies will be represented in this way, not all of which will be of primary interest to the user. Also, only a few of the fundamentals will be of interest. Thus, the totality of company data is largely redundant, and decisions are most easily made by monitoring key companies or entities, and within these, only certain fundamentals. The data items 606 represent user-selected companies or financial entities, from which image data is to be generated. The fundamentals used to generate image data may be defined separately for each of the possible views 801 to 806.

Figure 11

The grid view 801 indicated at 801 in Figure 8 has an appearance, when displayed on the monitor 304, as shown in Figure 11. Five separate windows are displayed, including the view window 1101. A title bar 1102 includes drop-down menus for controlling various aspects of the windows that are being displayed.

A file menu 1103 enables loading and saving of user preferences, stored as configurations. A configuration stores configurations for several windows, and for multiple views. An edit menu 1104 enables the user to modify the selection of data items currently used to generate display data. A views menu 1105 enables the user to switch between the available 3D views 801 to 806 for the selected data items. A window menu 1106 enables the user to edit items or characteristics in any of the windows that are currently active. A help menu 1107 provides context-sensitive help to enable the user to navigate the several windows and menus that are provided. An edit menu 1104 enables the user to set up equations for generating parameters in addition to those provided by the company fundamentals. Any of the fundamentals may be supplied as a variable to an equation defined by the user, the result of which may then be applied in the same way as a company variable when generating display data. A feed menu 1109 enables the user to configure aspects of the connection with the data service provider 101.

A data items window 1111 displays a list of currently selected data items that are being used to generate display data. Usually, more companies will be selected than can fit into the display, and the user may scroll through the list using a scroll bar or other scrolling means provided by the operating system 601. A share set window 1112 provides a short list of companies for which fundamentals are to be displayed. A single company selected in this window, for example MSFT, will be highlighted in some way in the view window 1101 , and its fundamentals will be displayed in a fundamentals window 1113.

A grid properties window 1114 enables the user to define which fundamentals are to be used for each dimension of the grid of data items that is displayed in the view window. In the present example, fourteen companies have been selected as data items from which display data is to be generated. These are arranged as elements within a four-by-four array. Each of the two dimensions of this array is associated with a particular fundamental, or user-defined value. The X dimension has its fundamental selected at 1121 , and the Z dimension has its fundamental selected at 1122. Examples shown are Market Cap and Dividend for X and Z dimensions respectively. However, any fundamental or user defined value may be used. The array is sorted so that companies have Market Cap values that increase in the X dimension, and Dividend values increase in the Z dimension. A third dimension, Y, is also defined, for which a different properties window is used.

Figure 12

A properties window for defining the Y dimension of the grid view 1101 shown in Figure 11 is shown in Figure 12. A weights properties window 1201 defines weighted fundamentals contributing to the Y dimension for each of the companies from which image data is to be generated. Three value fields 1211 , 1212 and 1213 are provided in which the user defines three characteristics applied to a company fundamental. These fields define the name of the fundamental 1221, whether or not to invert the fundamental 1222, and the level of weighting 1223 for the fundamental. In the simplest situation, only a single fundamental is used, in which case the weights in fields 1212 and 1213 are set to zero. The fundamental in field 1211 is multiplied by its weight 1223, and this value defines the extent of a company's height in the 3D bar graph shown in the view 1101. Inverting the fundamental, by ticking the box at 1222, has the effect of using the reciprocal of the fundamental, with the result multiplied by its weight 1223. A normalise values option 1224 causes the heights in the Y dimension, of each company on the grid, to be normalised in response to the height of the highest company, which is set to a value of one.

When multiple fundamentals are configured to contribute to the Y dimension of a company in the grid view 1101, this enables the user to set up sophisticated interrelations between fundamentals, that may be quickly and easily adjusted for maximum effectiveness when visualising company performance. Reconfiguration of the weights properties window 1201 and the grid properties window 1114 shown in Figure 11, enables the user to easily adapt and try out various methods for generating image data from company fundamentals.

In the grid view 1101 , or any of the 3D views, the user can modify the viewpoint in 3D space, so as to obtain a different perspective on the data items that have been rendered. For example, the viewpoint my be rotated left or right, up or down, with the grid of companies remaining at the centre. This may be achieved by moving the mouse cursor over the view window 1101 , and then dragging the mouse 303 with its left button held down, left or right, up or down, with respect to the screen. The user can zoom in or out of the view window by dragging the mouse, but this time with the control key held down on the keyboard 302 as well as the mouse button. In this way the user may quickly and easily view the data that is being displayed from a number of different 3D perspectives, without the complexity of having to navigate the 3D environment with a special 3D input device such as a 3D mouse or a dataglove. Commands for modifying the viewpoint of a view in 3D space are provided by OPENGL 3D graphics instructions. The user's viewpoint is specified as a single set of x, y, z co- ordinates and a 2D projection of the 3D objects viewed from that point in world space is rendered.

In any of the views, the user may identify a particular company in the show data window 1112, and this company will be highlighted in the 3D view by changing its colour to a contrasting colour.

Figure 13

The planets view 802 indicated in Figure 8, is shown in Figure 13. The planets view window 1301 has a circular grid 1302. Each company or data item is represented as a planet orbiting the centre 1304 of the grid

1302. The centre is marked by a small yellow sphere 1304. The radius, or size, of a planet 1303 is controlled by a first company fundamental selected in the planet properties window 1305 in a first data entry field 1306. The distance of a planet 1303 from the first circle 1315 of the grid 1302 is determined by a second company fundamental selected in a second data entry field 1307. A reference point 1308 on the outer circumference of the grid 1302 provides a reference for a zero degree angular displacement. The reference point is also marked by a small yellow sphere. The angular displacement of a planet 1303 from the reference point 1308, around the grid, is controlled by a third company fundamental selected in a third data entry field 1309.

Each data entry field 1306, 1307 and 1309 includes a region 1310 for selecting the company fundamental and a region 1311 for a scaling factor that may be adjusted in accordance with the requirements of the company data from which image data is to be generated. The circular grid, the planets and the reference point rotate smoothly about the centre 1304 at a fixed rate, typically once every two seconds. The rate of rotation is set by a rotation parameter at 1312 and may be adjusted by activation of the "X" and "Y" keys of the keyboard 302. Rotation enables the user to see the relationships and characteristics of data items from a continuously changing perspective. Because the orbit paradigm is used, the continuous rotation of perspective does not have the disconcerting effect it would have if applied to data items on non-circular grid. The vertical perspective position may be adjusted by dragging with the mouse in a vertical direction in the view window 1301.

Figure 14 The nails view 803 indicated in Figure 8 is detailed in Figure 14. The nails view 1400 is similar to the planets view shown in Figure 13. Each data item is represented by a shaft with a cap, having an appearance like that of a nail. In this view four company fundamentals are used to generate image data. The radius, distance and angle 1401 , 1402 and 1403 are defined in the same way as the radius, distance and angle in the planets view 802, except that the radius relates to the radius of the cap of each shaft. An additional fourth company fundamental is used to control the height of the cap above the plane of the circular grid. This fourth fundamental is selected at data entry field 1404. As with the planets view, the rate of rotation of the circular grid, and all the objects displayed on it, can be adjusted. In both the planets view and the nails view, rotation may be paused by activation of the "p" key of the keyboard 302.

Figure 15 The particles view 804, indicated in Figure 8, is illustrated in Figure

15. In the particles view, data items are shown as uniform sized spherical particles. These are initially located randomly in a volume of 3D space. Each particle, representing a data item, has three contributing company fundamentals, defined in data entry fields 1501, 1502 and 1503 in the particles properties window 1504. Each fundamental has an associated factor, that determines its contribution to the behaviour of each data item. The particles view window 1505 shows a model in which three attractors 1511 , 1512 and 1513 exert a gravitational force of attraction upon each data item. The attractors are red, blue and green in color. The magnitude of the attraction is dependent upon the inverse square law of gravitation, and upon the mass of the data item with respect to red, blue and green dimensions of gravity, as defined at fields 1501 , 1502 and 1503. Thus, for each company item, three gravitational forces are identified, with respect to each of the red 1511 , blue 1512 and green 1513 attractors. A vector addition is performed, followed by an integration with respect to time, in order to determine a movement of a company item 1521 in 3D world space.

Figure 16

Dynamic or tertiary conditions may provide indications of market behaviour. After a period of time, some or all of the data items will have moved to a stationary position, or a condition of localised positional oscillation between two of the attractors. This is illustrated in Figure 16. During the simulation, it is possible for trajectories to take data items outside of the bounding volume of the model. As a data item, or particle, hits the inside of the bounding 3D volume, it is reflected, and thereby returned to a region of space where it may be influenced by the pull of one or all of the three attractors. In the planets, nails and particles view the objects used to represent companies in 3D space may be color coded in accordance with their market sector. A typical colour scheme is as follows: Basic Materials - Red Capital Goods - Olive Green Conglomerates - Light Green Consumer Cyclical - Dull Yellow Consumer Non-Cyclical - Light Purple

Energy - Cyan Financial - Dark Grey Healthcare - Light Grey Services - Dark Purple Technology - Gold

Transportation - Dark Cyan Utilities - Orange

The provision of this uniform colour scheme across several switchable views enhances the coherency, quality and quantity of information created by browsing the data.

Figure 17

The top list view 805, indicated in Figure 8 is shown in Figure 17. The list view differs from the other views in that the user-defined data requirement 704, shown in Figure 7, defines the number of data items to display, but does not select them individually. Selection of individual data items is performed automatically during the process of generating image data. The user may, however, at step 704, define an overall collection of data items from which data items will be automatically selected. An example of this is a user definition of a requirement to select US stocks, shares and options.

In the top list view window 1701 a plurality of lists is displayed. Each list is sorted in accordance with a particular company fundamental or user- defined type. The same company appears at different positions in each list. By clicking on a data item 1702 in a particular list, the data item turns blue in colour. The same data item is highlighted by a blue colour 1703 in its position in other lists. However, because the user defines and thereby restricts the number of items in the lists, it is possible that the company will not appear in all of the lists that are shown in the window 1701. When a company is highlighted in this way, its details are displayed in the fundamentals window 1707. The user defines the number of items that appear in each list in the items dialog 1704 in the top list properties window

1705. The user adds or deletes fundamentals from a list of lists at 1706. Any number of lists can be displayed in the top list window 1701 , and this number is limited only by the clarity and resolution of the monitor 304. Similarly the limit on the number of list items 1704 is limited by clarity. Each data item, represented by a column in each list in the top list view 1701 , is colour coded. A white column indicates that the position of the data item within the list has remained unchanged since the last data update. A green column indicates that a data item is rising in value with respect to the average value of items in the list. A red column indicates that a data item is decreasing in value with respect to the list average. A gold- coloured column indicates that the element has just appeared in the visible part of the list. Each column has a plus (+) or minus (-) sign on its top surface. This indicates whether the absolute value of the data item is rising or falling.

Figure 18

The process of generating the grid view shown in Figures 11 and 12 is summarised in Figure 18. A company and its fundamentals are considered as individual data items 1801, 1802, 1803. The company is illustrated by a name followed by numerical fields. At least the three numerical fields are required. Figure 18 shows an example in which fourteen companies, A to N, are used to generate image data. It is intended to compare the third numerical fields of all fourteen companies. The companies, A to N, are indexed as elements in a two dimensional array 1811. The initial location of data items A to N in the array 1811 is not important. A sorting process is performed in which two-by-two areas 1812 of the array are selected and analysed. In each step of analysis, elements of the selected two-by-two area are compared to see if a swap is necessary.

The analysis procedure is illustrated at 1821 to 1824. At 1821 , the horizontal dimension of the array is considered. The lower two elements of the two-by-two area 1812 are examined. Their first numerical fields 1831 are compared. If the first numerical field 1831 is larger in the rightmost element, then this is considered as being the correct order, as it increases from left to right. Otherwise, a swap is performed between the elements, so that correct left right ordering is achieved. Correct left right ordering is considered with respect to the first numerical field 1831. At 1822, the two leftmost elements are examined. Their second numerical fields 1832 are compared. In this case, appropriate ordering is considered as being an increase from the bottom to the top of the array 1811 , and a swap between these elements is performed, if necessary, to achieve correct ordering in this dimension. At 1823, the horizontal dimension of the array is considered again.

The upper two elements of the two-by-two area 1812 are examined. Their first numerical fields 1831 are compared, and if the first numerical field 1831 is smaller in the rightmost element, then a swap is performed between the elements. At 1824, the two rightmost elements are examined. Their second numerical fields 1832 are compared. In this case, a swap is performed to ensure that the ordering increases from the bottom to the top of the array 1811 with respect to the second numerical field 1832. Each two-by-two area 1812 selected from the array 1811 overlaps the previous selected area. Thus, proceeding horizontally, the first area is ABEF, the second BCFG and so on. When proceeding vertically, the first is ABEF, the second EFIJ, and so on. The array is fully scanned a row at a time, with each row beginning ABEF, EFIJ and so on. If, at any point in scanning the array, a swap is necessary, this enforces the condition that another complete scan must be done, once the present one is finished. Several scans will usually be completed before the array becomes stable, and no more swaps occur. Eventually, the result is an array 1813 that has been sorted in both its dimensions, with respect to first and second numerical fields of its elements.

Having performed this sorting procedure, the third numerical fields 1833 are used to define an extension in a third dimension for each data item, as shown at 1814. Preferably a plurality of contributing numerical fields is combined in a weighted sum in order to define this third dimension for each data item. The array may then be rendered using standard 3D rendering instructions provided by the OPENGL 3D graphical environment.

As a result of this process, companies A to N are in the closest possible proximity, while but are ordered with respect to two numerical fields. Additional numerical values from each company's data, are used to define the third dimension for the view of the resulting sorted array. By generating image data in this way, valuable information is created. The first, second and third numerical fields may be defined as any of the company fundamentals, or a plurality of company fundamentals may be combined to define the third dimension for each element in the array. Companies may be displayed in this way at the centre of a regular two-dimensional grid projected into a three dimensional view. Their proximity and spatial relationships provide immediate and valuable information to the user.

Figure 19

The steps required to implement the grid view are summarised in Figure 19. At step 1901 the number of data items selected by the user at 702 is counted, and assigned to a variable N. At step 1902 a variable W is defined as being the integer portion of the square root of N. At step 1903 a variable H is defined as being the integer portion of one plus the square root of N, which is the same as W + 1. At step 1904 a two dimensional grid array G is defined having a width of W elements and a height of H elements. W and H only differ by one, and this arrangement guarantees that all data items will fit into a roughly square grid array. At step 1905 the grid array G is populated with data item pointers. Superfluous grid elements are supplied with pointers to a data item having zero in each of its numerical fields. At step 1906 the grid is sorted, as illustrated in Figure 18. At step 1907 the grid is rendered in a 3D projection, as shown in Figures 11, 12 and 18.

Figure 20

The step 1906 of sorting the grid, shown in Figure 19, is detailed in Figure 20. At step 2001 a variable F is set to zero. At step 2002 the first or next value for X is selected. X is a variable that is used to scan the width of the grid array G. It takes possible values from 1 to W - 1. At step 2003 the first or next value for Z is selected. Z is a variable that is used to scan the height of the grid array G, and takes possible values from 1 to H - 1. At step 2004 a two-by-two sub-array of elements is selected in the grid using X and Z as the base value. Comparisons are made between these elements, with swapping performed if necessary. At step 2005 a question is asked as to whether a swap occurred at step 2004. If so, F is set to the value one. At step 2007 a question is asked as to whether another Z value is required. If so, control is directed back to step 2003. Alternatively control is directed to step 2008, where a question is asked as to whether any X values remain for selection. If so, control is directed to step 2002. Alternatively, all the grid G will have been scanned. At step 2009 a question is asked as to whether F has the value zero. If so, this indicates that an entire scan of all the elements in the array has been performed without a swap occurring. This condition indicates that the array G has been entirely sorted in both its X and Z dimensions, and therefore the sort is complete. Alternatively, if the value of F is one, as set at step 2006, the grid is scanned again, swapping as necessary, until no more swapping occurs. Eventually, at step 2009, the value of F will be zero, and this completes the grid sort.

Figure 21 The step 2004 of comparing elements of a two-by-two grid, shown in

Figure 20, is detailed in Figure 21. The steps of Figure 21 are also summarised visually in Figure 18 at 1821, 1822, 1823 and 1824. At step 2101 a comparison is made between the first numerical fields of data items referenced by grid elements at G[X,Z] and G[X+1 ,Z]. It is intended to sort elements by their first numerical fields in the X dimension of the array, increasing with increasing values of X. Thus, if elements are found to be in the wrong order in this respect, at step 2102, then they are swapped at step 2103. If a swap occurs, this completes the steps of Figure 21. At step 2104 a comparison is made between second numerical fields of data items referenced by grid elements at G[X,Z] and G[X,Z+1]. It is intended to sort elements by their second numerical fields with increasing values of Z. If incorrect ordering is detected at step 2105, then a swap is made at step 2106. The same tests are performed in steps 2107 to 2112 with respect to a comparison between elements horizontally at G[X,Z+1] and G[X+1 ,Z+1], and finally vertically again between elements at G[X+1 ,Z] and G[X+1,Z+1].

Swapping of elements in the grid is performed in response to a comparison of data that are indexed by grid elements, but not actually stored in an area of physical memory 402 that is used to store a grid. In most instances, data is sorted via pointers stored as elements in the grid, and those pointers themselves may point to other pointers, and so on, eventually leading to the actual numerical value. In this way, the amount of data movement within physical memory is kept to a minimum. It will be understood, however, that in many instances it is convenient to talk of swapping or moving data in memory or in an array, when in fact this is referring to a swap or move of pointers.

Figure 22 The step 1907 of rendering the grid in a 3D projection, shown in

Figure 19, is detailed in Figure 22. At step 2201 the first or next W value for the grid G is selected. W values are integers that correspond to each of the columns of elements across the width of the array. At step 2202 the first or next H value is selected. H values are integers that correspond to each of the rows of elements in the grid array G. At step 2203 the data item indexed by a pointer stored at element G[W,H] in the grid is identified. At step 2204 the values for W and H set at steps 2201 and 2202 are used to define X and Y positions of a vertical column in the 3D grid view 1101. At step 2205 the height of the vertical column is calculated as a Z value. This value, Z, is determined by combining a plurality of weighted numerical fields, fundamentals, from the data item indexed at G[X,Z]. At step 2206 a column is drawn representing the data item, using values for X, Y and Z determined at steps 2204 and 2205. OPENGL drawing commands are used. At steps 2207 and 2208 remaining values for H and W are selected. Steps 2201 to 2208 select and draw each company selected by the user in response to a sorted grid of companies, and weighted fundamentals that define the height of columns above this grid, in a 3D projected view 1101.

Figure 23

The step 2205 of calculating a Z value from wieghted numerical fields, shown in Figure 22, is detailed in Figure 22. The steps of Figure 23 will be understood with reference also to Figure 12. At step 2301 variables Q, R and S are assigned to fundamentals defined as values one, two and three in fields 1211 , 1212 and 1213 in Figure 12. At step 2302 variables A, B and C are assigned to respective weights for values one to three, also defined in fields 1211, 1212 and 1213. At step 2303 a question is asked as to whether it is necessary to invert value one. This is answered with respect to a tick box 1222 for the respective field 1211 , 1212 or 1213. If the box

1222 is ticked, then the value for Q is inverted at step 2304. A similar check is made at steps 2305 to 2308 for the remaining respective fields. Finally, at step 2309, an equation combines variables A, B, C, Q, R and S to generate a value for the variable Z, that is used in the rendering of a column for a data item at step 2206.

Claims

Claims

1. Data processing apparatus, including memory means for storing instructions and data items, and processing means for generating image data from a plurality of said data items in response to said stored instructions, wherein said data items have at least four numerical fields, including a first numerical field, a second numerical field and a plurality of contributing numerical fields; said processing means performs repeated operations of; defining a position for each of said data items by processing said first numerical field and said second numerical field to generate a first position co-ordinate and a second position co-ordinate; defining a visual characteristic for each of said data items by combining weighted sums of said contributing numerical fields; and rendering image data representing said data items in a three dimensional view by combining said first co-ordinate and said second coordinate with said visual characteristic.

2. Apparatus according to claim 1 , therein said processing means performs a rendering step such that said visual characteristic is height.

3. Apparatus according to claim 1, wherein said numerical fields represent company fundamentals.

4. Apparatus according to claim 1, including reception means for receiving data from a data service provider, wherein said data items are a user-selected subset of data received from said data service provider.

5. Apparatus according to claim 1, wherein said processing means is configured to generate first and second position co-ordinates by performing operations of: defining a two dimensional array in which said data items are elements; followed by repeated steps of: swapping elements in said array in its first dimension in response to a comparison of values in said first numerical fields; and swapping elements in said array in its second dimension in response to a comparison of values in said second numerical fields.

6. Apparatus according to claim 5, wherein said processing means is configured to repeat said swapping operation until a condition is attained to the effect that: a first numerical field of elements in rows of said array are arranged in ascending order; and the second numerical fields of elements in columns of said array are arranged in ascending order.

7. Apparatus according to claim 1, wherein said processing means is configured to receive user identified weights for fundamentals of respective companies or financial entities and to define said weighted sums in response to said user identified weights.

8. Apparatus according to claim 1, wherein said processing means is configured to invert a fundamental in response to a user request for an inverse weighted fundamental.

9. Apparatus according to claim 1 , including a network and receiving means for receiving data from said network wherein said data items are generated in response to data received via said receiving means.

10. Apparatus according to claim 1, including a network configured to supply updated data, wherein the said image data is generated in response to updated data received over said network and includes modified data items from which image data has been previously generated.

11. A method of generating image data from a plurality of data items, in a data processing system including memory means for storing instructions and said data items, and processing means for generating said image data; said instructions defining operations to be performed by said processing means to generate said image data from said data items; said data items having at least four numerical fields, including first and second numerical fields, and a plurality of contributing numerical fields; wherein said operations include repeated steps of: defining a position for each of said data items by processing said first and second numerical fields to generate first and second position coordinates; defining a visual characteristic for each said data items by combining weighted sums of said contributing numerical fields; and rendering image data representing said data items in a three dimensional view by combining said first and second co-ordinates with said visual characteristic.

12. A method according to claim 11, wherein said visual characteristic is height.

13. A method according to claim 11, wherein said numerical fields are company fundamentals.

14. A method according to claim 11 wherein said data items are a user-selected subset of data received from a data service provider.

15. A method according to claim 11 , wherein said first and second position co-ordinates are generated by steps of: defining a two dimensional array in which said data items are elements; subsequent repeated steps of: swapping elements in said array in its first dimension in response to a comparison of values in said first numerical fields; and swapping elements in said array in its second dimension in response to a comparison of values in said second numerical fields.

16. A method according to claim 15, wherein said swapping steps are repeated until the following condition is attained: the first numerical fields of elements in rows of said array are arranged in ascending order; and the second numerical fields of elements in columns of said array are arranged in ascending order.

17. A method according to claim 11, wherein said weighted sums are defined by user-identified weights for fundamentals of respective companies or financial entities.

18. A method according to claim 11, including inverting a fundamental in response to a user request for an inverse weighted fundamental.

19. A method according to claim 11, wherein said data items are generated from data received over a network.

20. A method according to claim 11, wherein said image data is generated in response to a data update received over a network that includes modified data items from which image data has previously been generated.

21. A computer system programmed to execute stored instructions such that in response to said instructions said system is configured to perform methods according to any of claims 11 to 20.

22. A computer-readable medium having computer-readable instructions executable by a computer such that, when executing said instructions, a computer will perform the method steps according to any of claims 11 to 20.

23. A computer-readable memory system having a plurality of data fields stored therein representing a data structure, wherein said data structure facilitates the execution of a method according to any of claims 11 to 20.

24. In a computer system having a graphical user interface including a display and a user interface selection device, a method of providing and selecting from a menu on the display facilitated by method steps according to any of claims 11 to 20.

25. A method of generating image data from a plurality of data items substantially as herein described with reference to the accompanying drawings.