- Tytuł:
-
Z problematyki porównywalności kartogramów
Some problems of choropleth maps comparability - Autorzy:
- Leonowicz, A.
- Powiązania:
- https://bibliotekanauki.pl/articles/204291.pdf
- Data publikacji:
- 2002
- Wydawca:
- Polskie Towarzystwo Geograficzne
- Tematy:
-
kartografia
kartogram
mapa
porównywalność - Opis:
-
Artykuł jest próbą odpowiedzi na pytanie, jakie warunki powinny spełniać kartogramy, aby możliwe było poprawne ich porównywanie. Przedstawione czynniki. które mogą mieć wpływ na wizualną ocenę podobieństwa tych map.
Process of choropleth map comparison bases on two elements: space units and class intervals. Basing on them one can formulate conditions of optimaln comparability: similar pattern of space units on both maps, the same number of classes, the same and method of class selection, The question of class selection, which would provide for comparable maps has not been finally solved yet. In the selection of class intervals several hints can be followed: 1) class interval system with approximately equal nymbers of observations in each category should be chosen, which prevents empty classes in case of different statistical distribution; 2) division methods should base on statistic measures treating each statistical sets as a whole, rather than on absolute values, different for each set; 3) a unified measure should be used for describing class limit values, regardless on the denomination of mapped data. These conditions are best met by two types of class selection methods: arithmetic mean and quantilies is the first; the second bases on standard deviation units. Their advantages are combined in the method proposed by A. Dąbrowski (1980). His division bases on arithmetic mean and quantiles, with class limits modified so that they can be set as standard deviation units. The method of M.S. Monmonier (1975) is an attempt to optimise the class selection methods for choropleth map comparison. It sets class limits at intervals, which stress similarities between maps. Effects of classification method on the correlation of presented phenomena has been researched by J.Olson (1972a, 1972b). In the research of comparability it is important to consider map perception by users. Process of visual comparison of choropleth maps has been researched in many experiments (H.H.McCarty, N.E. Salisbury 1961; J. C. Muller 1976a, 1980; R. Lloyd, T. Steinke 1976, 1981, 1983; M. P. Peterson 1979). Their results show, that visual judgement of map similarity does not always confirm to the correlation calculated for particular maps. This process is not fully random. Among the factors, which shape user perception are the following: similar distribution of phenomena presented on maps, relative blackness, complexity, similarity between map patterns, correlation rate (positive or negative), degree of association (certain levels of cerrelation may be more difficult to judge), user's experience in map reading. Further, space units can gain relative significance due to their size, shape and location (central, peripheral), belonging to a class (extreme or middle) and geographical knowledge of the map reader. The choice of class selection method can influence the map reception by affecting some of the above factors. Application of equal area method, for example, makes it possible to obtain similar blackness on all maps, and therefore eliminates this factor's influence on the judgement of similarity. It is more difficult to limit the influence of other factors - so far there are no other such solutions. Attempts to reach comparability of choropleth maps meet many obstacles. One should then consider more efficient ways of presenting relations between phenomena. Two-variable choropleth map can be such a way, although it should be noted that it is also much more difficult to interpret. Therefore it seems necessary to conduct further research on maps, which would facilitate correct evaluation of relations between presented phenomena. - Źródło:
-
Polski Przegląd Kartograficzny; 2002, T. 34, nr 1, 1; 22-33
0324-8321 - Pojawia się w:
- Polski Przegląd Kartograficzny
- Dostawca treści:
- Biblioteka Nauki