Statistics Toolbox
Expert tools
Statistics Toolbox
The statistics toolbox has visualisation and analytical functions
The visualisation includes the Data visualiser. The DV is written in JAVA and can be used to visualise multidimensional tables stored in GESMES XML format. The user selects an indicator from a list of available indicators and two of the available dimensions (these will be the x and y axis of a two dimensional table). Furthermore the user selects a specific value for the remaining classifications. The DV, then presents a two dimensional table with the requested values.
The analytical functions of the Statistics toolbox fall in three categories
The first is descriptive statistics where simple statistics and plots will be generated to accompany the maps.
The second is data mining techniques that will provide for: § Clustering: Following a user query we need to determine clusters of the data objects (regions) based on vectors of thematic data and then depiction them in the STATLAS mapping system. § Classification: Define a set of thematic information and given some user-selected training set of representative regions to classify the remaining regions according to one or more appropriate classification methods. § Spatial Data Mining. Perform the above tasks (clustering and classification) and include position and topological information (like distance, adjusted to etc) to find associated regions that share some thematic features. This should include all geographical features like roads, waterways etc.
The third will be a full range of application of methods termed as spatial econometrics. § Explore Spatial Relationships: Determine if some values are spatially correlated using coefficients like the Moran Coefficient and the Geary ratio that have known asymptotic distributional properties and can therefore be used for hypothesis testing. Produce Moran Coefficient Scatterplots. Explore the need and feasibility of multivariate spatial autocorrelation. Compute the sample variogram and the variogram cloud. § Model spatial correlations by means of a spatial autoregressive process therefore enhance statistical description and improve the referential basis for statistical decision-making. Furthermore, provide for the prediction of missing values.
The Statistics Toolbox is at the stage of implementation and testing by a team consisting of Statistics and IT personnel from Liaison Systems and NTUA.
Digital Terrain Model [up]
Provide an overview of the result which gives the reader an immediate impression of the nature of the result, its relevance and its potential; Briefly describe the current status/applications of the result (if appropriate) with non confidential information on entities potentially involved.
The Digital
Terrain Model (DTM) is a representation of the earth's surface displayed
as grey-scale hill shading. The DTM can be displayed as background to
the mapped statistical data to help orient the user within the map space,
as well as for purely aesthetic reasons. At the moment, we have used the
GTOPO30 data to create a DTM with a resolution of 30 arc seconds, that
is approximately one kilometer. We have compiled a DTM that covers the
furthest extents of the European Union, as well as possible future member
states.
Since the DTM has a z-axis, it could potentially be used to develop a
3D model of the topographic features of the European Union. A 3D model
would be useful for creating panoramic views of selected areas, or for
creating perspective views with statistical data "draped" over
the DTM. Arbitrary perspective views of the surface can be generated,
thus supplementing the usual perpendicular view of maps. 3D models will
further be used as a basis for the display of thematic information that
is closely related to topography.
Expert tools [up]
For object-oriented analysis in a GIS and Multimedia Cartography environment, 2D and 3D graphics are interrelated with a structured data base and a set of dedicated GIS functions. Expert tools allow customised data presentation, further exploration of statistical maps in conjunction with 3D topographic models. These expert tools are developed to be used easily but yet to provide a maximum of visual feedback. They can be used interchangeably for both statistic and topographic data, classification and colour change, or modification of the illumination of 3D models. The expert tools are developed in the framework of the Multimedia Maps Extension part of the system (also called imap). The functionality of those tools are as follows: 2D data analysis, regional queries, change analysis. Once again, the development will lead to a a set of software functions which will altogether form a toolbox. The main innovation is the ability to exchange such modules in this projects but also in further projects of the involved institutions. At the moment, this result and especially under those premises, must also be considered research oriented and experimental. We aim at a high interoperability but can only prove this in follow-up projects.
