KARTOGRAFISCH TIJDSCHRIFT
van kartografische hypothesen: de
computergesteunde kaartgebruiks-
test. Kartografisch Tijdschrift xvi,
nr. pp. 44-48.
[6] Mersey, J.E. (1990), Color and the-
matic map design. Cartographica
37, nr. 3, pp. 1-157.
[7] MacEachren, A.M. (1982), The
role of complexity and symboliza-
tion method in thematic map ef-
fectiveness. Annais of the Associa
tion of American Geographers 72, nr.
4. PP- 495-513-
[8] Gilmartin, P. E. Shelton (1989),
Choropleth maps on high resolu-
tion crt's. The effect of number of
classes and hue on communication.
Cartographica 26, nr. 2, pp. 40-52.
[9] Mak, K. M.R.C. Coulson
(1991), Map-user response to com-
puter-generated choropleth maps:
comparitive experiments in Classifi
cation and symbolisation. Carto-
graphy and Geographie Information
Systems 18, nr. 2, pp. 109-124.
[10]Müller, J.-C. (1979), Perception of
continuously shaded maps. Annais
of the Association of American Geo
graphers 69, nr. 2, pp. 240-249.
[uJBrassel, K.E. J.J Utano (1979),
Design strategies for continuous-
tone area mapping. The American
Cartographer 6, nr. 1, pp. 39-50.
[12]Depuydt, F. (1988), Optimization
of the cartographic communication
process in choropleths by Compu
ter assistance. International Year-
book of Cartography 28, pp. 203-211.
[13]Declercq, F.A.N. (1995), Choro
pleth map accuracy and the num
ber of class intervals. In: Procee-
dings International Cartographic
Conference (17a: 1995, Barcelona),
pp. 918-922, Barcelona.
[14]Declercq, F. (1995), Visualization
of environmental data using gis
and desktop mapping: pitfalls and
success. In: Harnos Z. (ed.), Pro-
ceedings of the International Confe
rence on Environment and Informa-
tics, June 29 - July 1, 1995, pp. 38-
45. Hungarian Academy of Scien
ces, Budapest.
[15]Ormeling, F.J. (1992), Core concepts
in cartographic communication. Pro-
ceedings of the Seminar on Sur-
veying and Mapping Aspects of Vi
sualization and gis, pp. 27-35.
Utrecht: ica Commission on Edu-
cation and Training.
I996-XXII-4
Noot
De heer F.A.N. Declercq is vorig jaar aan de Katholieke Universiteit
te Leuven gepromoveerd. Dit artikel is naar aanleiding daarvan ge-
schreven. De oorspronkelijke titel van de (niet uitgegeven) dissertatie
luidt Computergestuurde visualisatie van kaartinformatie: interpolatie,
klassenindeling, symboolgebruik. Prof. F.L.M.L.J. Depuydt was de pro-
motor van het onderzoek.
Summary
F.A.N. Declercq F.L.M.L.J. Depuydt - More efficient Classi
fication and colour use in choropleth maps
Keywords: visualisation, thematic cartography, research
The efficiency by which map users can extract quantitative in-
formation from choropleth maps is assessed. By means of a map
use test, the influence of the type of colour scheme (area fill),
the number of classes and the rounding of the class limits are
investigated. Results show that the use ofonly gray tones prohi-
bits the reading of map information compared to the use of
schemes containing more contrasting variations of hatch pat
tern or colour. Choropleth maps with fewer than 6 classes are
not suited to display local map information: the use ofy to 11
classes is hereby recommended. To accomplish a perceptually ef
ficient presentation of regional variations the use of9 to 11 clas
ses becomes too complex, whereas again the 7 class maps yield
optimal scores. Map users capture map information more pre-
cisely and faster if class limits are rounded to one digit instead
of two or three digits afier the decimalpoint.
Resume
F.A.N. Declercq dr F.L.M.L.J. Depuydt - Classification et
emploi de couleurs plus efficaces pour des cartes choropletes
Mots-cles: visualisation, cartographie thematique, recherche
L'ejficacite avec laquelle les utilisateurs de cartes choroplethes
peuvent extraire de l'information quantitative et ordonnee est
examinee. Partant d'un test d'emploi des cartes, Tinfluence du
type de variable retinienne (implantation zonale), du nombre
de classes et de la fagon d'arrondir les bornes de classes est eta-
bli. Les resultats montrent que l'usage d'une Variation de va-
leurs de gris donne une lisibilite profondement inferieure com-
pare aux progressions contenant des variations de couleurs ou
de semis. Des choroplethes constituant de moins de 6 classes ne
sont pas propre pour lire des informations locales dans la carte:
l'usage de 7 11 classes est recommende pour ceci. Pour accom-
plir une repartition en struetures regionales par une maniere
efficace, l'emploi de 9 11 classes devient trop complex, tandis
que les cartes 7 classes nouveau obtiennent un score opti
mal. L'extraction d'information se fizit plus precisement et plus
vite si les bornes de classes sont arrondies jusqu'ä un chiffre au
lieu de deux ou trois chiffres.
29