## R.: Visualising changes in fund manager holdings in two and a half-dimensions (2004)

Venue: | Information Visualization |

Citations: | 16 - 1 self |

### BibTeX

@ARTICLE{David04r.:visualising,

author = {David and Mr. Tim Dwyer and Dr. David and R. Gallagher and R. Gallagher and Tim Dwyer and Tim Dwyer and Key Words},

title = {R.: Visualising changes in fund manager holdings in two and a half-dimensions},

journal = {Information Visualization},

year = {2004},

volume = {3},

pages = {227--244}

}

### OpenURL

### Abstract

Abstract. We explore a multiple-view, or overview and detail, method for visualising a high dimensional portfolio holdings data-set with attributes that change over time. The method employs techniques from multidimensional scaling and graph visualisation to find a two dimensional mapping for high dimensional data. In both the overview and detail views, time is mapped to the third dimension providing a two and a half dimensional view of changes in the data. We demonstrate the utility of the paradigm with a prototype system for visualisation of movements within a large set of UK fund managers ’ stock portfolios.

### Citations

686 |
Readings in information visualization: Using vision to think
- Card, Mackinlay, et al.
- 1999
(Show Context)
Citation Context ... visual form is going to challenge their “perceptual bandwidth”. The approach we present decomposes this visual challenge into two sub-problems, providing us with a basis for an “overview-and-detail” =-=[6]-=- visualisation design: • to visualise the changing holdings of all fund manager portfolios in our data-set in order to obtain a broad picture of the entire market; • to obtain a detailed view of chang... |

518 | A Heuristic for Graph Drawing
- Eades
- 1984
(Show Context)
Citation Context ...ics by introducing extra geometric parameters. For example, Brodbeck et al. [5] use such metrics to create a large graph which is then arranged in the plane using a force directed placement algorithm =-=[10, 14, 15, 23]-=-. In this way more highly correlated commodities are positioned geometrically more closely together such that clustering is clearly visible. Gross et al. [17] use a similar technique to compare econom... |

392 |
Modern Multidimensional Scaling: Theory and Applications
- BORG, GROENEN
- 2005
(Show Context)
Citation Context ... at a particular point in time. Obviously, to make an intelligible visualisation this high-dimensional space must be reduced to two or three dimensions. This is the domain of Multidimensional Scaling =-=[1]-=-, the chief aim of which is to find a lower dimensional representation for a high dimensional data-set. This representation should preserve, as much as possible, the relative Euclidean distances betwe... |

102 | A Fast Adaptive Layout Algorithm for Undirected Graphs
- Frick, Ludwig, et al.
- 1994
(Show Context)
Citation Context ...ics by introducing extra geometric parameters. For example, Brodbeck et al. [5] use such metrics to create a large graph which is then arranged in the plane using a force directed placement algorithm =-=[10, 14, 15, 23]-=-. In this way more highly correlated commodities are positioned geometrically more closely together such that clustering is clearly visible. Gross et al. [17] use a similar technique to compare econom... |

96 |
Preferences for stock characteristics as revealed by mutual fund portfolio holdings
- Falkenstein
- 1996
(Show Context)
Citation Context ...echnical aspects of portfolio management to investors who are not familiar and/or sophisticated in understanding quantitative portfolio analysis. For more information about performance analytics, see =-=[13, 22, 4]-=-. The work we present here is an ambitious attempt to visualise very high dimensional data. We explore a visualisation of the changing ownership interests of thousands of fund managers that aims to pr... |

95 |
Applied multivariate data analysis
- Everitt, Dunn
- 2001
(Show Context)
Citation Context ...here the dimensionality of the data is significantly lower than the number of data points it is more practical to use a classical multidimensional scaling method based on Principal Component Analysis =-=[12]-=-. In PCA the complexity is based on the dimensionality of the data rather than the number of points. PCA aims to find the axes of greatest variance (principal components) through our m-dimensional dat... |

88 |
How to draw a directed graph
- Eades, Sugiyama
- 1990
(Show Context)
Citation Context ...term used to describe rock consisting of layers of sediment. We use this term to avoid confusion with the conventional usage of the term “layered graph drawing” associated with the Sugiyama algorithm =-=[11]-=-. Visualisations of Stratified graphs have appeared in the literature from time to time, but the question of how to arrange them is yet to be thoroughly examined. Koike [19] was possibly the earliest,... |

53 | Visual Unrolling of Network Evolution and the Analysis of Dynamic Discourse
- Brandes, Corman
(Show Context)
Citation Context ...sibly the earliest, using a stratified graph to show the flow of control in parallel software execution, but the author left placement of the columns representing vertices to the user. Brandes et al. =-=[2]-=- chose to represent social group interactions as stratified graphs using the well known forcedirected method simulating physical attractive and repulsive forces in an iterative algorithm that attempts... |

26 | Visualizing related metabolic pathways in two and a half dimensions
- Brandes, Dwyer, et al.
- 2004
(Show Context)
Citation Context ...ta such that the layout is optimised for stratified graphs. More detail of the modification is provided in our description of a metabolic pathway visualisation based on stratified graph visualisation =-=[3]-=-. The 2 1 2D column view system is based on the WilmaScope graph visualisation engine, which is freely available, open sourced and downloadable from http:// wilma.sourceforge.net/. WilmaScope was crea... |

23 | Domesticating Bead: Adapting an Information Visualization System to a Financial Institution
- Brodbeck, Chalmers, et al.
- 1997
(Show Context)
Citation Context ...n abstract concepts. The most common examples have tried to capture stock price correlation based on various similarity metrics by introducing extra geometric parameters. For example, Brodbeck et al. =-=[5]-=- use such metrics to create a large graph which is then arranged in the plane using a force directed placement algorithm [10, 14, 15, 23]. In this way more highly correlated commodities are positioned... |

11 | Visualising a fund manager flow graph with columns and worms
- Dwyer, Eades
- 2002
(Show Context)
Citation Context ...nsions are used to represent more complex dimensional scaling or graph layout. The work we present here expands on ideas originally presented in conference papers. The detailed view was introduced in =-=[9]-=- and the overview discussed in [8]. In this extended paper we expand greatly on both ideas, particularly the overview has been refined based on feedback from expert users as discussed in Sections 4 an... |

4 | A scalable method for visualising changes in portfolio data
- Dwyer
- 2003
(Show Context)
Citation Context ...complex dimensional scaling or graph layout. The work we present here expands on ideas originally presented in conference papers. The detailed view was introduced in [9] and the overview discussed in =-=[8]-=-. In this extended paper we expand greatly on both ideas, particularly the overview has been refined based on feedback from expert users as discussed in Sections 4 and 8. 4. Overview Visualisation usi... |

1 |
Active investment manager portfolios and preferences for stock characteristics: Australian evidence
- Brands, Gallagher, et al.
- 2004
(Show Context)
Citation Context ...echnical aspects of portfolio management to investors who are not familiar and/or sophisticated in understanding quantitative portfolio analysis. For more information about performance analytics, see =-=[13, 22, 4]-=-. The work we present here is an ambitious attempt to visualise very high dimensional data. We explore a visualisation of the changing ownership interests of thousands of fund managers that aims to pr... |

1 |
Factor analysis: an introduction to essentials
- Cattell
- 1965
(Show Context)
Citation Context ... associated with the two largest eigenvalues to define the projection plane. This, by definition, captures the greatest variance in the data. For example, Figure 5 gives a histogram, or scree diagram =-=[7]-=-, of the eigenvalues for the components or eigenvectors of the trivial synthetic data matrix used in Figure 4. However, one can just as easily use any pair of eigenvectors. Allowing the user to choose... |