WO2003079232A2

WO2003079232A2 - Credit data visualisation system and method

Info

Publication number: WO2003079232A2
Application number: PCT/NZ2003/000048
Authority: WO
Inventors: Thomas Richard Beard
Original assignee: Compudigm International Limited
Priority date: 2002-03-18
Filing date: 2003-03-18
Publication date: 2003-09-25
Also published as: AU2003214726A8; WO2003079232A3; AU2003214726A1

Abstract

The invention provides a data visualisation system and method particularly suitable for visualising complex data associated with credit risk analysis. The credit data visualisation system comprises a ratings database of ratings data maintained in computer memory representing credit ratings of respective organisations and a retrieval device configured to retrieve from the ratings database credit ratings data of a plurality of organisations. The system includes a data constructor configured to construct a set of discrete data values from the retrieved data and a display configured to display respective graphical representations of the discrete data values. The invention further provides a related method of visualising credit data.

Description

CREDIT DATA VISUALISATION SYSTEM AND METHOD

FIELD OF INVENTION

The invention relates to a data visualisation system and method, particularly suitable for visualising complex data associated with credit risk analysis.

BACKGROUND TO INVENTION

Analysing credit risk requires access to and analysis of large volumes of complex data. At any given time, each issuer, being a company that issues credit instruments such as bonds, is assigned credit ratings by various rating agencies for example Moodys or Standard & Poors. An important aspect of credit risk analysis is tracking how issuers move from one rating category to another, and especially to detect those that default on credit or are removed from the market for other reasons.

A standard technique of credit risk analysis involves a rating transition matrix. The matrix tabulates the rate at which issuers move from one rating to another over a given time period. Such matrices are useful, but since they are numeric they are hard to understand and do not show patterns in a compelling way.

Furthermore, the rating transition matrix does not facilitate assessing how credit risk varies across different markets and industries. Each of these dimensions may be studied individually, but it is most enlightening to look at both simultaneously. Standard tools such as tables, bar charts or pie charts cannot show sufficient data to enable such analysis. SUMMARY OF INVENTION

In one form the invention comprises a credit data visualisation system comprising a ratings database of ratings data maintained in computer memory representing credit ratings of respective organisations; a retrieval device configured to retrieve from the ratings database credit ratings data of a plurality of organisations; a data constructor configured to construct a set of discrete data values from the retrieved data; and a display configured to display respective graphical representations of the discrete data values.

In another form the invention comprises a method of visualising credit data comprising the steps of maintaining in computer memory a ratings database of ratings data representing credit ratings of respective organisations; retrieving from the ratings database credit ratings data of a plurality of organisations; constructing a set of discrete data values from the retrieved data; and displaying respective graphical representations of the discrete data values.

BRIEF DESCRIPTION OF FIGURES

Preferred forms of the credit data visualisation system and method will now be described with reference to the accompanying Figures in which: Figure 1 shows a block diagram of a system in which one form of the invention may be implemented;

Figure 2 shows the preferred system architecture of hardware on which the present invention may be implemented;

Figure 3 shows preferred form data structures in accordance with the invention; Figure 4 is one preferred form display generated in accordance with the invention;

Figure 5 is a flow chart of a preferred form look-up function of the invention; Figure 6 is a variation on the display of Figure 4;

Figure 7 shows further preferred form data structures in accordance with the invention;

Figure 8 shows a further preferred form display generated in accordance with invention;

Figure 9 shows a further preferred form circular representation generated in accordance with invention; and Figure 10 shows a flow chart of a preferred form of the invention.

DETAILED DESCRIPTION OF PREFERRED FORMS

Figure 1 illustrates a block diagram of a preferred system 10 in which one form of the invention may be implemented. The system includes one or more clients 20, for example 20A, 20B and 20C, which each may comprise a personal computer or workstation described below.

One or more of the clients could be interfaced to a workstation 30 as shown in Figure 1 via a network or collection of networks 40. The networks could include a local area network, a wide area network, a WAP connection, the Internet, or any combination of the foregoing. The preferred system 10 may include a data repository 50, for example a data warehouse maintained in computer memory. It is envisaged that the data repository may comprise a single database, a collection of databases, or a data mart. The preferred data repository 50 includes data from a variety of sources and may include, for example, credit ratings data 52, industry code data 54 and/or location data 56.

Credit ratings data 52 represent the credit rating assigned to one or more issuer organisations. A credit ratings agency such as Standard & Poors or Moodys assigns a credit rating to an issuer organisation. The rating is one selected from a plurality of rating classes.

Standard & Poors issues ratings in one of 18 rating classes from class 1 indicating the best rating to class 18 indicating a "default" or worst rating. These 18 rating classes are referred as AAA, AA+, AA, AA-, A+, A, A-, BBB+, BBB, BBB-, BB+, BB, BB-, B+, B, B-, CCC and D respectively. The ratings are preferably ordered so that ratings deteriorate from class 1 to class 18.

As an alternative data structure, the credit ratings data 52 could include eight distinct rating classes referred to as AAA, AA, A, BBB, BB, B, CCC and D respectively. An organisation within one of these rating classes could optionally be assigned a "+" indicator or "-" indicator to show relative credit rating within a rating class.

An organisation is assigned an initial rating class for example between AAA and CCC. Depending on the performance of the organisation and other factors, the organisation could be assigned a higher class in what is known as a credit upgrade or could be assigned a lower class known as a credit downgrade, hi some circumstances, an issuer may end up in default in which it is assigned a "default" rating class referred to as D or could end up without a rating and has the attribute "not rated" or N .

The industry code data 54 contains a plurality of industry codes. An organisation is associated with an industry code that indicates the type of market in which the organisation operates. One example of industry codes conforms to the standard industrial classifications "SIC" index. The SIC index is separated into 11 divisions labeled division A to division K. Each division may in turn be further divided into a plurality of classes. Division A could include agriculture, forestry and fishing and cover classes 01 to 09, division B could cover mining and be divided into classes 10-14, division C could cover construction covering classes 15-17 and so on, with division K covering non-classifiable establishments covering class 99.

It is envisaged that any suitable classification system could be used. One example is the North American industry classification system "NAICS" that has replaced the Standard Industrial Classification for some data sets, although there are several data sets still available with SIC based data. A further alternative could be the Australian and New Zealand industrial classification (ANZSIC) that is a standard classification used in Australia and New Zealand for collection, compilation and publication of statistics by industry.

It is envisaged that the industry code data base 54 be accessed directly from an official source containing a plurality of organisations with corresponding industry codes, or alternatively the industry code database could comprise a database containing organisation data associated with industry code data compiled from one or more of such official sources.

Location data 56 could include region and country indicators for various organisations. Examples of region indicators could include Asia/Pacific, Europe, Latin America, North America and country indicators could include China, Spain, Mexico or Canada for example. Each organisation could be associated with one or more region indicators and one or more country indicators available from an official source. These region and country indicators could include major markets in which an orgamsation has commercial interest, or could indicate the domicile or home country/region of an organisation. The location database 56 could represent an official source or alternatively could represent data compiled from one or more official sources.

The workstation 30 could include a data memory 32 and server 34 and preferably operates under the control of appropriate operating and application software stored in and running on data memory 32.

The system 10 includes a retrieval device or retrieval component 60 which in one form comprises a software-implemented query enabling retrieval of data from the data repository 50, the data memory 32 and/or a data memory forming part of a client 20. The retrieval device 60 could alternatively comprise a data memory in which such a query is installed and operating.

Data retrieved with the retrieval device 60 is processed with the server 34. A data constructor 70 is configured to construct a set of discrete data values from the data retrieved from the retrieval device 60. The system 10 could also include an aggregator 80 configured to optionally construct a set of summary data values from two or more of the discrete data values constructed from the data constructor 70. Both the data constructor 70 and aggregator 80 are preferably implemented as software routines or software programs and are further described below.

A display 90 in one form comprises a computer implemented software program that is configured to generate a display of the data obtained from the retrieval device 60, data constructor 70 and/or aggregator 80 on a client workstation 20, as will be described below. The display could alternatively comprise a conventional display software module interfaced and operable to display data on the client 20.

The display 90 is configured to display respective graphical representations of summary data values generated by the aggregator 80. Individual data values are preferably displayed in a colour and/or location on a data display device that is representative of the size of the data value. It is envisaged that colour scale table 92 provide a look up table to allocate certain colours to data value ranges. Similarly, a position look up table 94 provides X and Y coordinates on a data display device for displaying a representation depending on the information that the data value represents and the particular display of which the representation forms a part.

It will be appreciated that the system of Figure 1 is conceptual in nature. In one form a client 20 could have installed on it credit ratings data 52, industry code data 54, location data 56, the retrieval device 60, the data constructor 70, the aggregator 80 and the display 90 and function independently of any other component in the system. Alternatively, some or all of the components could be installed and operating on the workstation 30 and the client access required information as necessary in a client server relationship.

Figure 2 shows the preferred system architecture of a client 20 or workstation 30. The computer system 200 typically comprises a central processor 202, a main memory 204, for example RAM, and an input/output controller 206. The computer system 200 may also comprise peripherals including a data entry device such as a keyboard 208, a pointing device 210 for example a mouse, trackball, touchball or cursor key set, a display or screen device 212, a mass storage memory 214 for example a hard disk, floppy disk or optical disk, and an output device 216 for example a printer. The system 200 could also include a network interface card or controller 218 and/or a modem 220. The individual components of the system 200 could communicate through a system bus 222.

It is envisaged that the invention have a wide area of application and that the nature and format of the data stored in computer memory will be different for each application. Different applications of the invention are set out below, hi one form the invention is configured to display data representing credit risk data for a set of organisations.

Figure 3 illustrates preferred form data structures generated in accordance with the invention. Typically, an organisation will be assigned a rating by a rating organisation and that rating will be effective as of a defined date and/or time. The retrieval device in accordance with the invention is configured to retrieve from computer memory, for example the credit ratings database 52, industry code database 54, location database 56, data repository 50, data memory 32 and/or client 20, data representing credit ratings and worthiness of one or more organisations. An example of a collection of such data is shown at 300.

The retrieved data could be stored in a data table that includes issuer identifier 302, rating class identifier 304 and a date/time indicator 306 representative of the date/time at which a rating class is assigned to an issuer.

The retrieval device preferably stores the retrieved data 300 in computer memory for subsequent retrieval. The data constructor then constructs a set of discrete data values from the retrieved data.

One example of the resulting data is shown at 330. The data could include an issuer identifier 332, a rating class identifier 334 assigned to an issuer at a first time interval (rating^ and a rating class identifier 336 assigned to an issuer at a second time value (rating₂).

It is envisaged that the time value associated with credit rating 336 be greater than the time value associated with credit rating 334. The preferred time interval between the two time values is in the range of three years to five years.

The data constructor is preferably configured to store data 330 in computer memory ready for subsequent retrieval. As an alternative, it is envisaged that data 330 could be retrieved, compiled and stored in computer memory by the retrieval device. The data 330 represents credit ratings at two different time intervals for a plurality of issuer organisations.

In one form the data constructor is configured to construct a set of discrete data values from the data 330. Typical data values constructed by the data constructor are illustrated at 340. Data 340 could represent a rating transition matrix for a set of issuers over a predefined time period. The data could include a credit rating 342 at a first time interval (ratingi), a second credit rating 344 representing the credit rating of an organisation at the end of a three or five year time period (rating₂), and a frequency value 346.

The frequency value 346 represents the frequency of issuers that have been assigned initial credit rating 342 and that have been assigned credit rating 344 at the end of the time period. The frequency indicator could be represented as a percentage of issuers in a set of issuers, a proportion of issuers in the set of issuers or the number of issuers.

Figure 4 shows one preferred representation generated in accordance with the invention. In one form, the display is configured to generate a grid 400 of individual graphical representations of the discrete data values 340 from Figure 3. On one axis of the grid, the display positions credit rating at an initial time period indicated at 402. Credit rating classes could range from AAA down to CCC.

Positioned along another axis of the grid is the credit rating data at the end of the three or five year time period indicated at 404. The rating classes range from AAA down to CCC and also include default class D and not rated class NR.

The frequency of organisations having an initial credit rating 402 and a final credit rating 404 are presented to a user as a representation positioned in a location within representation 400. For example, a representation of the frequency of organisations that have been assigned an initial rating class of A+ and that have been downgraded to A are represented by a graphical representation centred at 406.

In one preferred form the display is configured to access a colour scale look up table 92 shown in Figure 1 in order to highlight features of interest within representation 400. Referring to Figure 4, representations indicated generally at 408 are assigned the colour red, issuers indicated generally at 410 are assigned the colour orange, issuers indicated generally at 412 are assigned the colour green and issuers indicated generally at 414 are assigned the colour grey. Initial and final credit rating combinations for which issue of frequencies are 0 are shown in white.

Figure 5 illustrates a flow chart of the operation of the look up table. In one preferred method 500, an initial ratingi is retrieved 502 from computer memory. Rating₂ after a three or five year period is also retrieved 504 from computer memory.

In some cases the frequency of issuers having ratingi and rating₂ values will be 0. If this test 506 is satisfied, the colour of the representation is set 508 to a suitable background colour for example white.

Rating₂ is then tested 510 as to whether rating has a "default" or "not rated" credit value. If so, the representation is set 512 to red.

If the value of rating₂ is less than the value of ratingi indicated at 514 the value of the representation is set to orange. Rating₂ will be less than ratingi in circumstances where an issuer or set of issuers has been subjected to a credit downgrade.

If rating₂ is greater than ratingi as indicated at 518 the issuer or a set of issuers has experienced a credit upgrade and the representation is set 520 to green.

If rating is equal to ratingi as indicated at 522, the issuer or set of issuers has experienced neither a credit upgrade nor a credit downgrade and has retained the same credit rating. The representation colour is set 524 to grey.

Where there are further rating₂ values to retrieve from computer memory 526 these are retrieved successively. Where there are further ratingi values to retrieve from computer memory 528 these are retrieved successively from computer memory 528.

In this way graphical representations can be displayed in one of a plurality of colours based on the credit rating at an initial time value and a credit rating at a later time value. In a further preferred form, the display could be configured to assign an intensity value for the representation based on the frequency value. In one implementation of the invention the frequency value could be used to retrieve an intensity value from a look up table indexed by frequency value or range of frequency values.

The colour codes make it easier to distinguish issuers who are improving credit ratings shown in green, from issuers whose credit ratings are deteriorating shown in orange. Issuers who default or leave the market are shown in red to draw particular attention to these issuers, and issuers who remain on the same rating are shown in grey.

It is also envisaged that the display has access to a position table 94 from Figure 1 indicating the position in the representation 400 from Figure 4 in which a particular representation should be positioned. The position table could include and be indexed by a set of (ratingi, rating ) combinations together with an X and Y coordinate position at which the corresponding representation should be displayed. Alternatively, the X and Y positions of each representation could be generated by a function using the rating values as inputs to the function.

The visualisation of Figure 4 has several advantages over numerical presentation. Volatility in the market is shown visually in which stable markets cluster near the diagonal, while volatile markets tend to spread over the representation. Broad patterns are made obvious. The balance between risers and fallers is shown more clearly. Anomalies are highlighted such as the high proportion of issuers making a transition in Figure 4 from AA- to B indicated at 420. Special emphasis is also given to issuers that default or leave the market indicated at 408.

Figure 6 shows an improvement to the representation of Figure 4 in which the representation of discrete data values could include contoured representations, as is more particularly described in WO 00/77682 published 21 December 2000. Rather than a simple geometric square, the representations include contoured representations positioned at predefined locations within a representation that in some cases can be easier to view than a mosaic of squares. The benefits of this representation are most readily apparent when animated.

Figure 7 illustrates further preferred forms of data retrieved by the retrieval device. The data 700 could include ratings data similar to the ratings data 300 shown in Figure 3. The data could also include industry code data 720 comprising a plurality of issuer codes 722 with corresponding industry codes 724.

The data 700 could also include location data 730 comprising a plurality of issuer codes 732 and region codes 734 and/or country codes 736. The retrieval device preferably stores this retrieved data 700 in computer memory for subsequent retrieval.

The retrieval device could also generate and store data 740. This collection of data 740 could include issuer codes 742, a rating at initial time value 744 (ratingi), a rating at a second time value three or five years later at 746 (rating₂), a location identifier 748 and industry code identifier 750.

The data constructor is preferably configured to construct and store resulting data for example data 760. This data could include a location identifier 762, an industry code identifier 764 and a frequency identifier 766.

The discrete frequency values 766 could represent the cumulative default rate, the number of issuers from the set of issuers that has a default rating at the end of the time period. The discrete values 766 could alternatively show company count, marginal default rate, cumulative default rate, weighted average rating, spread of ratings and rating trends.

Figure 8 illustrates a sample representation that the display could generate from data 760 from Figure 7. The representation 800 could comprise a grid of individual representations of respective data values positioned based on the industry code data and/or the location data. In one example region indicators are positioned on one axis of a grid for example 802. The regions could include North America, Europe, Asia, Africa and Latin America. Within each of these regions, the Y coordinates over which each region is displayed could be further divided into individual country Y coordinates.

Along another axis of the grid could be positioned industry code data for example 804. The industry codes could be grouped into a series of categories, each category spanning a range of X coordinates. Each category could be further divided into individual classification codes and each classification code assigned a unique X coordinate.

Individual representations of the discrete values from Figure 7 could be positioned in the representation 800 based on region and industry code data. The individual representations could be shown as squares or alternatively could be represented as contoured representations.

Individual data values could be assigned a colour using a colour look up table and could be assigned a position in representation 800 using a position look up table. Alternatively, the colour and/or position could be calculated by one or more functions using data values, location code, values and/or industry code values as inputs.

Figure 9 illustrates a further preferred form representation that could be generated from the data of Figure 7. Individual graphical representations of data values are positioned in a substantially circular configuration based on the industry code data and/or the credit ratings data.

Data associated with a particular credit rating is positioned along a radius from centre point 902 based on credit ratings. For example, organisations with a credit rating CCC could be positioned at radius 904 whereas organisations with a AAA rating could be positioned at radius 906 that is larger than radius 904. Credit ratings between CCC and AAA could be positioned between radius 904 and radius 906. The representation 900 is also preferably divided into a plurality of sectors based on industry code data. For example, the industry classification "capital" could be positioned in the sector 910 whereas metals could be positioned in the sector 912.

Each industry group could be further divided into more specific industry categories. The food category 914 for example could be further divided into a plurality of categories, some of these categories indicated in expanded view 920.

It is envisaged that the display could have access to a position table that, given as input a rating class and industry code, selects the appropriate position within representation 900 to position the representation. Alternatively, the position in X and Y coordinates could be selected as a result of a function taking as input an industry category and a rating class.

The individual representations could be contoured as shown in Figure 9. The contours could show any of a range of key performance indicators such as marginal default rate, cumulative default rate, volatility or spread. In the case of an analyst looking at a particular portfolio, the key performance indicator could represent the value of that portfolio, split among various issuers invested in, and showing where risk is increasing or decreasing over time.

The display could also make use of a colour scale table that could provide background shading to indicate rating bands. Credit ratings between AAA and A- could be rated as green, ratings between BB+ and B- could be rated as yellow, and a rating of CCC could be indicated as red to show risk.

The representations of Figures 8 and 9 in particular illustrate how credit risk can vary across different industries and geographic markets, enabling such factors to be examined simultaneously.

Figure 10 illustrates one preferred method 1000 of operation of the invention. Data is retrieved 1010 from computer memory using a suitable query initiated by the retrieval device. The retrieved data could include credit ratings, industry code and/or location data.

A set of discrete data values is then constructed 1020 from the retrieved data. This set of data values could include for example the frequency or volume of issuers having a particular combination of ratingi and rating values.

A set of summary data values could then be optionally constructed 1030 from the set of data values. A representation of these discrete data values could then be displayed 1040 and a representation of the summary data values could also be displayed.

The foregoing describes the invention including preferred forms thereof. Alterations and modifications as will be obvious to those skilled in the art are intended to be incorporated within the scope hereof, as defined by the accompanying claims.

Claims

1. A credit data visualisation system comprising: a ratings database of ratings data maintained in computer memory representing credit ratings of respective organisations; a retrieval device configured to retrieve from the ratings database credit ratings data of a plurality of organisations; a data constructor configured to construct a set of discrete data values from the retrieved data; and a display configured to display respective graphical representations of the discrete data values.

2. A credit data visualisation system as claimed in claim 1 wherein the retrieval device is configured to retrieve from the ratings database credit ratings data of a plurality of organisations issued at a plurality of time intervals.

3. A credit data visualisation system as claimed in claim 2 wherein the set of discrete data values constructed by the data constructor includes an organisation identifier, a credit rating at a first time value and a credit rating at a second time value.

4. A credit data visualisation system as claimed in claim 3 wherein the set of discrete data values constructed by the data constructor includes the credit rating at the first time value, the credit rating at the second time value, and a frequency value.

5. A credit data visualisation system as claimed in claim 4 wherein the display is configured to display one or more of the discrete data values in one of a plurality of colours based on the credit rating at the first time value and the credit rating at the second time value.

6. A credit data visualisation system as claimed in claim 4 or claim 5 wherein the display is configured to display one or more of the discrete data values in one of a plurality of colour intensities based on the frequency value.

7. A credit data visualisation system as claimed in any one of the preceding claims, wherein the display is configured to display a grid of individual representations of respective data values.

8. A credit data visualisation system as claimed in claim 7 wherein the individual representations are positioned in the grid based on the credit ratings data.

9. A credit data visualisation system as claimed in any one of claims 1 to 6 further comprising industry code data maintained in computer memory representing industry codes of respective organisations, the retrieval device configured to retrieve industry code data from computer memory.

10. A credit data visualisation system as claimed in claim 9 further comprising location data maintained in computer memory representing the locations of respective organisations, the retrieval device configured to retrieve location data from computer memory.

11. A credit data visualisation system as claimed in claim 9 or claim 10 wherein the display is configured to display a grid of individual representations of respective data values, the individual representations positioned in the grid based on the industry code data.

12. A credit data visualisation system as claimed in claim 10 wherein the display is configured to display a grid of individual representations of respective data values, the individual representations positioned in the grid based on the location data and/or the industry code data.

13. A credit data visualisation system as claimed in claim 9 wherein the display is configured to display the individual graphical representations in a substantially circular configuration, the individual representations positioned in the circular configuration based on the industry code data.

14. A credit data visualisation system as claimed in claim 9 wherein the display is configured to display the individual graphical representations in a substantially circular configuration, the individual representations positioned in the circular configuration based on the industry code data and/or the credit ratings data.

15. A method of visualising credit data comprising the steps of: maintaining in computer memory a ratings database of ratings data representing credit ratings of respective organisations; retrieving from the ratings database credit ratings data of a plurality of organisations; constructing a set of discrete data values from the retrieved data; and displaying respective graphical representations of the discrete data values.

16. A method of visualising credit data as claimed in claim 15 further comprising the step of retrieving from the ratings database credit ratings data of a plurality of organisations issued at a plurality of time intervals.

17. A method of visualising credit data as claimed in claim 16 wherein the set of discrete values includes an organisation identifier, a credit rating at a first time value and a credit rating at a second time value.

18. A method for visualising credit data as claimed in claim 16 or 17 wherein the set of discrete values includes a credit rating at a first time value, a credit rating at a second time value, and a frequency value.

19. A method of visualising credit data as claimed in claim 17 further comprising the step of displaying one or more of the discrete data values in one of a plurality of colours based on the credit rating at the first time value and the credit rating at the second time value.

20. A method of visualising credit data as claimed in claim 18 further comprising the step of displaying one or more of the discrete data values and one of plurality of colour intensities based on the frequency value.

21. A method of visualising credit data as claimed in any one of claims 15 to 20 further comprising the step of displaying a grid of individual representations of respective data values based on the credit ratings data.

22. A method of visualising credit data as claimed in any one of claims 15 to 20 further comprising the steps of maintaining in computer memory industry code data representing industry codes of respective organisations; and retrieving industry code data from computer memory.

23. A method of visualising credit data as claimed in claim 22 further comprising the steps of maintaining in computer memory location data representing the locations of respective organisations; and retrieving location data from computer memory.

24. A method of visualising credit data as claimed in claim 22 or claim 23 further comprising the step of displaying a grid of individual representations of respective data values, the individual representations positioned in the grid based on the industry code data.

25. A method of visualising credit data as claimed in claim 23 further comprising the step of displaying a grid of individual representations of respective data values, the individual representations positioned in the grid based on the location data and/or the industry code data.

26. A method of visualising credit data as claimed in claim 22 further comprising the step of displaying the individual graphical representations in a substantially circular configuration based on the industry code data.

27. A method of visualising credit data as claimed in claim 22 further comprising the step of displaying the individual graphical representations in a substantially circular configuration based on the industry code data and/or the credit ratings data.