Choropleth map

Abstract: The correlation between a pair of distributions is often measured with the product-moment correlation coefficient, r. But the choropleth maps of the same distributions may be more alike or less alike than that coefficient would indicate. The degree to which map likeness corresponds to the correlation between the data before mapping is affected by the class interval method used. Using rank correlation (Kendall's t) to measure the map correlation and drawing data from normal distributions, it is found that: (1) there is a mathematically expressible curvilinear relationship between the product-moment correlation and the expected map correlation; (2) this relationship appears to remain identical regardless of class interval system; (3) the standard deviation about the expected values does vary from one class interval system to another. Smaller standard deviations are associated with increased numbers of observations within each variable and with increased numbers of intervals. For a given r value, number of o...

Abstract: In this paper we propose a solution to several key limitations of current web based mapping systems: slow rendering speeds and the restriction of online map viewing to a small number of areal units as well as a limited number of users. Our approach is implemented as a Scalable Tile Map Service that distributes dynamic choropleth maps in real-time through a new caching methodology. This new Map Service lays the foundation for advances in web based applications reliant on dynamic map rendering such as emergency management systems and interactive exploratory spatial data analysis. We present the results of an empirical illustration in which this new methodology is used to facilitate collaborative decision making by visualizing spatial outcomes of simulation results on the fly.

Abstract: The “Technological Transition” has had a tremendous impact on cartographic processes and outputs. Atlases, whether national, regional or urban, were affected as they moved from a static to an animated and interactive era. The overall positive development, consolidated by map automation and map animation, has taken various forms and directions. Nevertheless, some graphic and cartographic outputs, especially atlases, still suffer from serious pitfalls, regarding some semiological aspects of the resulting maps and graphics. The Digital Urban Atlas of Jeddah (DUAJ) is a research in progress aware of such deficiencies. It tries to spare it from some frequent errors in map design related to: the choice of the base map, the use of some frequent symbols such as columns and pie charts for multivariate quantitative data, and the alternative solutions to choropleth maps. These deficiencies are selected and examined among others. The authors are convinced that the generated maps and diagrams that should be produced in digital atlases must be examined with regards to the semiological rules that drew upon the theory of signs perception. They try to incorporate the basic principles of graphic semiology reviewed, corrected and adapted to the GIS requirements. This is also meant to avoid map noises and deficiencies. The goal is to attract future Digital Atlases producers towards developing higher interest in map look and design and users to grasp messages rapidly. Specifically, the use of GIS and linked spatial databases within the DUAJ should not pay attention only to data handling and crossing, but also, to the semiological features of the outputs and relevant visual communication. In its preliminary stage, the DUAJ should give answers not only to the three questions raised in this research but also to others not examined here. This is to prevent from incorrect, irrelevant or inadequate use of the cartographic tools and, following the analysis of concrete examples, to propose a set of recommendations in establishing maps and graphics. When necessary, the DUAJ GIS-based outputs are inserted in CAD or CAC systems to benefit from some their specific visual tricks and subtleties that yield better legibility and efficiency of communication for Atlases users as recommended by authors. This experience is discussed regarding its eventual contribution to better map reading through some examples of maps and graphics from the DUAJ project (in progress).

Abstract: A simple yet a valid question often asked by GIS users is: what is the “optimal” number of choropleth map classes for a given data set? This question is barely addressed in the literature, however. In this present work, a method is therefore proposed and named the “Weighted Number of Classes Index (WNCI)”. It proposes an optimisation approach according to which the resultant class break values are weighted based on the numbers of their occurrences within a set of classification runs. This is followed by normalising the total weight for each classification run; the classification run that has the highest total normalised weight is chosen, and its associated number of classes is considered comparatively the best option for the given data set. Using seven data sets, the results showed that the WNCI method rationale and performance appeared to be unaffected by the types of data sets used, because producing the highest WNCI value for each dataset is possible, regardless of the dataset values and distribution. Further enhancement of the WNCI was proposed from a digital implementation perspective.