Syntactic Characteristics and a Smart Construction
Mechanism forThematic Map Symbols
Fei Zhao1, Qingyun Du2, Fu Ren2 , Guizhi Wang3
1.School of Resource Environment and Earth Science，Yunnan University，Kunming 650091, China;
2.School of Resources and Environmental Science, Wuhan University, Wuhan 430079, China;
3.National Geomatics Center of China, Beijing 100830,China
Abstract. The construction of thematic map symbols is a very complex and intellectually
demanding process, but thematic symbols can be automatically generated and easily shared
on the web using the syntactic structure of semantic symbols. In this paper, the symbol types,
inner structure, and design pattern are discussed. A syntactic construction theory based on
letter (thematic maps primitive) - word (single thematic symbol) - sentence (combined
symbols or complex symbols) structure model is proposed to automate the construction of
thematic map symbols. As a result of this research, symbols can be defined using
cartographic primitives which are arranged according to syntactic principles. A semiotic
model and word-centered construction theory can be integrated into interactive cartography
as represented by the technology of Internet. Finally, the concepts and schema of this theory
are discussed, and some examples are presented based a web thematic cartographic system to
verify its utility.
Key words: syntax; thematic map; symbol; linguistics; web cartography; construction
1.
Introduction
With the development of Internet of Things (IOT), thematic cartography must be real-time
and intelligent for dynamic monitoring of sensor data (Iosifescu-Enescu et al., 2010).
Intelligent thematic cartography is evolving into a system for intelligent selection of
representational method (TIAN, 2007) instead of expression content determined based on an
expert system (HUA, 1993). Now, intelligent thematic cartography is focusing on dynamic
intelligent generation of thematic symbols. There are two approaches: one is to analyze the
elements of symbols (MA, 1995;Schnabel & Hurni, 2009), and dynamically constructing
symbols according to constitutive rules of given symbols; the other is to mine thematic
mapping knowledge from produced atlases (Li et al., 2009), and then selecting thematic
symbols based on case reasoning mechanisms from artificial intelligence. The former method
generates symbols from elements and constitutive rules and can extend thematic symbol
types. It is limited in that the incomplete constitutive elements and rules may mean that the
thematic symbols produced by this method do not conform to spatial cognition rules. The
latter method can produce thematic maps according to existing and excellent thematic maps
or atlas as similar cases. However, it cannot satisfy dynamic mapping needs in a distributed
environment, as it depends on existing symbol characters. Web2.0 users have much more
interactive environment (Boulos et al., 2010), and are more inclined to participatory spatial
information services (Jankowski, 2009). How to analyze cartographic data rapidly? How to
choose appropriate representation mode? How to design the corresponding symbols? Which
construction modes should be followed in order to generate dynamic thematic maps? These
questions are addressed in this paper.
Intelligent thematic cartography in web context must analyze the nature and syntactic
characteristics of symbols according to philosophical theory (ZHU, 1988;YU, 1995;Fernández,
2012). This paper introduces semiotics and linguistics into thematic cartography, analyzes
map symbol systems from the perspective of phonetics, semantics, syntax, abstract semiotic
models of thematic maps, and constructs a thematic symbolic framework centered on “word“.
These help define thematic symbols according to three attributes which are symbolic
elements, structure principles and word relations, and selects construction attributes
intelligently based on different data characteristics, to generate thematic maps automatically.
Discussing The design model of this mechanism is discussed, and a formal description of a
prototype system and its integrated application are detailed.
2.
Syntactic Model of Thematic Map Symbols
The linguistics of cartography can be divided into macroscopic and microcosmic levels.
Microcosmic linguistics mainly focuses on symbolic structural elements, meanings, relations,
inner symbols and layers, and construction mechanisms (DU, 2004). Further, map symbol
systems can be divided into following parts from elementary to advanced:
①Symbol construction elements (Primitive, Visual Variable, etc.)——Phoneme；
②Single thematic symbol——word；
③combined or complex symbols——sentence.
Phonemes as minimal phonetic units form words according to semantic structure, words
form complex symbols through syntax, and these further form symbol layers. Symbol layers
can form thematic maps when overlayed with other layers. A “Phoneme” forms a map symbol
system with a syntactic structure in a unified semiotic meaning framework. The model is a
three-dimensional space including symbolic primitives (phonetic), symbolic references
(semantic) and symbolic organization (Syntactic). (Figure 1)
Syntax
Dimension
Syntax
Construction
External
Combination
Internal
Syncopation
Phoneme
Synonymy
Semantics
Hyponymy
Dimension
Meronymy
Antonymy
Phoneme
Suprasegmental
Combination Phoneme
Phonetics
Dimension
Figure 1. Syntactic model of thematic map symbol
The Phonetics-dimension is divided into three levels based on different degrees of
abstraction:
① The Phoneme is the most basic distinctive feature among symbolic constitution elements.
Point and line which are controlled by six distinctive features can for symbolic phoneme. The
six distinctive features include size, shape, orientation, brightness, color and saturation;
②The Phoneme Combination: Phoneme can form “Word Morpheme” (primitives) and “NonWord Morpheme” (visual variables);
③ The Suprasegmental Phoneme is a combination phonetic feature. It describes symbol
elements in detail, such as the area under a graphic line of curvature.
The Semantics-dimension includes symbolic reference and semantic features. A thematic
symbol system focuses on the semantic relations of words; synonymy, hyponymy, meronymy
and antonymy. There are three levels in the syntax-dimension: internal syncopation of words,
external combination, and syntax construction. The internal syncopation of words refers to
phoneme sets; External combination are words forming phrases based on semantic relations;
Syntax construction consists of two meanings, one is repeated using structural principles
between words or words and phoneme sets based on semantics to form symbolic sentences
and are a kind of combined relationship; while secondary phonemes at the sentence level
represents a new feature.
Phonetics-dimension defines the phonemes of symbols, forming morphemes, and then
forming words. The construction of this dimension is controlled by the inherent constructive
features of symbolic phonemes. It does not consider semantics and syntax. The
semantics-dimension and syntax-dimension effect morpheme word formation and
combination of words into sentences. The common carrier of syntactic structure is the word.
A word is the most active part of the symbol system. It is both a component in abstract
language systems and a specific unit in a real symbol system. Through an investigation
of“active symbols”, researchers can grasp the regular word-formation rules of thematic
symbols, and can further combine them with specific data and context to research
pragmatically the process of word meaning. Therefore, it is important to place the word at the
core position, and form the word-centered models in a thematic symbolization framework
(Fig.2). The framework includes three macro levels : phonetic, semantics and syntax,
corresponding with the syntactic structure. Words and sentences are two material carriers of
symbols. These three macro levels have a comprehensive effect on the process of word
construction, and combine to generate sentences.
Figure 2. Word-centered model for thematic symbolization
3.
Thematic Map Symbol Construction Model based on Syntactic
Structure
According to syntactic structure of thematic symbols and symbolization framework in this
paper, we combined the design principles of thematic symbols, made Primitives (Primitives,
P), retinal variables (Visual variables, V) for ther basic symbol elements, forming phoneme
sets (Symbol Phoneme, Sp). and Operator sets (Structure Operator, Os) employed as symbol
structure principles to construct word sets (Symbol Word, Sw) based on phonemes
according to the structure of the symbol building operators. Finally,a sentence set(Symbol
Sentence, Ss) including combined or complex symbols is generated by the combination
between phonemes and words based on the Structure Operator.Thus, the Symbol
set(Symbol, S) can be described as formulae(1):
S = (Sp, Sw, Ss)
(1)
And in which, Sp= Os(P, V) Sw= Os(Sp, Sp) Ss= Os(Sp, Sw)
This can be divided into three: the definition of symbols phoneme collection (Sp),
construction of symbols word collection (Sw) and combination of symbols sentence
collection(Ss).
3.1. Symbol Phoneme and its Combination
As mentioned, a symbol phoneme integrates basic primitives and visual variables. Its
combination can form all morphemes(minimal meaningful units) or rich syllables(still
meaningless units) at the phonetics level based on certain rules, such as Aesthetics, or gestalt
laws.Morpheme level symbolic elements are extracted to study the characteristics of the
word-building ability themselves apart from phoneme combinations which also have a great
deal of flexibility. So this section is aimed at determining primitives and the corresponding
visual variables with different word-building abilities to form phoneme sets.
There are different views on the basic primitive types (JIANG, 1986;LI & ZHOU, 1997;
Schnabel & Hurni, 2009). The first two authors aim to generate thematic symbols with single
combinations, but neglect the mapping relation between data and primitives or symbols.
Although the latter represent thematic data through the control of the scale of the statistical
symbol data direction, it is limited to the ideographical expression at the carto-word level
regardless of intelligent combination at syntactic level. Considering these facts, we extract
thematic symbol morphemes from a phoneme combination forming primitives and analyze
their cartographic characteristics, such as position, structure, data mapping and so on.
Based above and with reference of Primitive-based Construction Theory (Schnabel & Hurni,
2009),the Symbol Phoneme set as discussed in this paper can be abstracted as the phoneme
and its basic combination or three-dimensional form shown in Figure 3. The special
construction properties marked red in Figure 3 include a reference point,which indicates the
position in the map,and size or orientation of different dimensions.
A fixed radius is the construction property of points. A point can be used in constructing dot
scatter and coin charts when acquiring quantitative information by accumulating points or
identifying distribution density of points. The method for three-dimensional forms is the
same as described previously. Point primitives can be combined with other primitives to
generate a new symbol word controlled by visual variables. In linear charts, for example,
point primitives with different shapes are distinct other thematic types.
When describing thematic phenomenon with curves or polylines, the degree of symbol,
thickness is used to indicate the difference of quality grade; the symbol shape is formed by
color and linear combination. When describing thematic factors with statistical symbols, a
reference line must be defined. A series of points with vertical distance from reference line
can be linked to form a linear chart. Vertical distance is used to represent thematic data,
while lines connecting points become line segments or curves can form polyline or curve
charts respectively. Their three-dimensional form is used for statistical chart construction.
There are unidirectional and bidirectional types of flow, useful for presenting flow features
between two thematic phenomena such as population immigration and emigration. The
construction property for flow primitives is such that flow width is used to indicate certain
thematic data characteristics. An arrow direction is used to present movement direction and
color can distinguish quality features.
In the following primitives, the radius, semi-axis, inner and outer diameter, inner radius and
initial angle, and the inner and outer diameter and initial angle bear the construction
properties of circle, ellipse, ring, pie sector and ring sector, respectively. These primitives
generally employ radius as data direction and use other properties as the construction
direction. The three-dimensional form usually has a thickness property to represent data
direction, such as bar charts.
Triangle primitives include regular and right-angled triangles, so their construction
properties are side length, side length of right angle, and its angle, respectively. The side
length is data direction. The construction property of rectangle is height and width. When
applying both height and width to data representation, the area of rectangle is useful for
indicating data features and its reference point locates in the center of the rectangle. When
employing height for data representation and using width as a schematic expression, the
rectangle is common bar primitive. Its reference point locates the center of the base.
Additionally, its three-dimensional form generally applies to bar symbol or coin charts.
Regular polygons can use a number of sides or the distance from apex to center to define the
construction property; the latter is the data representation direction. Irregular polygons
generally are used either in special scales for thematic factors or abstract contour planes.
Thier construction property is controlled by a visual viable; brightness presents quantity
features and color or mode presents quality features. An irregular polygon generates
topological maps through area topological construction, and the construction property is in
the syntax dimension.
Fig.3 phoneme and its combination of thematic map symbol
The phoneme primitives in Fig.3 already have a certain visual variable properties at the level
of Symbol Phoneme set, such as outline width or color, filling color or pattern,etc..According
to the construction properties, primitives can represent each dimension of thematic data by
deflection change,size scaling, etc..But the semanteme of symbols or thematic data
contents can not be realized at the phoneme level; therefore a discussion on the
construction principles at the word level is in order.
3.2. Construction Principles in Word Level
On the basis of analysis of the Symbol Phoneme set construction properties, Symbol Word
(Sw) can be constructed with Sp following below Structure Operator(Os). We borrow and
extend ideas from current diagram construction theory (Bertin, 1983;Wilkinson et al., 2006;
Schnabel & Hurni, 2009).As shown in Figure 4, the black solid line indicates the data
mapping direction of Sp and the dashed one indicates the construction direction:
1. Point:Each spi in Sp constructs a word with itself located by the reference point.The
construction property representing thematic data usually is mapped with a statistical
indicator (Indicator, Ii)by Size, such as height of bar symbol,radius of circle, side length of
regular polygon and so on.It is usually used for proportional symbols bound to no word
morpheme(shape, color etc.)(Figure 4a) and can be used with correlative construction
principles integrally with the addition of a series of Operators(Rotation, Reflection etc.) as
described as formulae(2):
Os.Point(spi.position,spi.size(Ii),spi.color.Hue(Ii),spi.rotation,spi.reflection)
(2)
2. Linear:Multiple spi(sp1,…, spn) located by the reference point in Sp construct a word along
linear direction with specific interval. Aword is used in the visualization of two or more
dimensional data values,such as bar, stacked bar, dynamic circle
and line chart
etc..According to the relation between Sp construction direction and data mapping direction,
it can be divided
into there different concrete modes: parallel, orthogonal and
angled(correspond to Figure 4b, c and d).This construction principle has a strong linear
directivity so that it cab be used when the spi has regular pattern in a certain direction.Here
we denote the relation type of the direction between construction and data mapping by
Linear.direction, the interval of spi by Linear.interval.So it can be described as formulae(3):
Os.Linear(Linear.direction,Linear.interval…,Os.Point(spi.position,spi.size(Ii),
spi.color.Hue(Ii)),…)
(3)
3.Polar: Multiple spi(sp1,…, spn) located by the reference point in Sp construct a word along
polar direction with specific rotation angle.The spi with polar coordinates properties can
directly constructed by this principle based on their center.Others need auxiliary elements(a
circle generally) to make themselves have the defined properties. In this instance, it is used in
the visualization of two or more dimensional data values,such as circle, ring, sector and wing
chart etc..In this principle,the Sp construction direction is polar, meanwhile,the data mapping
direction can be divided into: polar and circular elevation (correspond to Figure 4e and
f).Here we denote the region of thematic data by Polar.direction, the value of spi by
Region.value and the num of spi by Region. num.Then the num of each row or column of
spi and the distances themselves can be calculated by Region.area and Region.value and can
be described as formulae(4):
Os.Polar(Polar.direction,Polar.startAngle,…,Os.Point(spi.position,spi.size(Ii),spi.color.Hue
(4)
(Ii)),…,Polar.angle, Polar.refCircle)
4. Region:Multiple spi(sp1,…, spn) in Sp construct word in a certain space(Grid, Region etc.)
repeated permutation by row-column(Figure 4g).Each spi represents a certain equivalent
value and the total number of primitives represents the visualized data value so an important
means for compensating the shortage of quantitative expression of other symbols. Several
groups represented each different value can be separately constructed by a Region then by
Line(using color to distinguish) to represent multi-dimensional thematic data,such as coin
charts, dot chart(the space is thematic region).Here we denote the region of thematic data by
Region.area(Grid or Polygon), the value of spi by Region.value and the num of spi by
Region. num.Then the num of each row or column of spi and the distance themselves can
be calculated by Region.area and Region.value and described as formulae(5):
Os.Region(Region.area, Region.value, Region.num,Os.Point(spi.position,spi.size(Ii), spi.color.
(5)
Hue (Ii)))
5. Topological:Certain allophones and word morpheme in Sp are constructed based on
topological relations, mainly representing cartograms, isograms, choropleths and flows. In
acartogram, the general topological relation is unchangeable in Sp, but a change in the
geometry of Sp, will realize representation between thematic data and topological distance or
topological area; Interpolating a contour based on discrete Sp objects to generate isogram; in
order to get isogram, we should use thematic data represented by a polygon phoneme itself to
match different color or mode after classification; flow is aimed at motion property of
thematic phenomenon to represent the flow condition between two sites. It is constructed by
selecting unidirectional or bidirectional streamline primitives based on motion path or
motion direction and maintaining topological relations based on its real or schematic and
other flow lines. The construction core for this method is the existing mature theory and
applications. For example, the distance topology catogram construction generation algorithm
(Shimizu & Inoue, 2009),the diffusion-based method for producing density-equalizing maps
generation algorithm (Gastner & Newman, 2004),the genetic algorithms-based equivalent
classification algorithm (Armstrong et al., 2003) and the graph partition-based multilayers
flow mapping generation algorithm (Guo, 2009). This paper will not focus on these kinds of
algorithms, but use the corresponding algorithms to realize construction of thematic maps.
Distance between primitives
Value per
primitive
Number of
primitives
per row
Fig.4 Construction principles in word level
3.3. Combination of syntactic dimension
Symbolic phoneme set Sp based on structural principles of letter level can construct a word
set Sw with referential meaning. Accordingly, Sw with referential meaning itself and semantic
relations between words can construct sentence with Os., and convey much more complex
thematic information. Aiming at different sematic relations between words, there are four
categories as following:
1. Synonym combination: refers to a kind of combination between words which can convey
same or similar thematic elements. A set of thematic elements can form various expressions
through construction principles and retinal variables decoration, then form an aggregation
system with all the representational possibilities. In this system, the combination with any
two words can form a kind of new express syntax. Different combinations with words can
stress different thematic content from different views. Take the resident population at
year-end as an example. There are three ways to express quantity. One is by choosing a
rectangular phoneme to form a columnar symbol in linear mode. The other is by choosing a
polygon phoneme based on topological construction to form a classification map. The last is a
combination of the two previous exprssions of quantity. Allocate columnar symbols to the
classification map and then convey the thematic elements with ratio scale and interval scale.
Generally, population is normalized by area to form a classification map of population
density, a synonym combination.
Two or multiple thematic elements with same or similar meaning content can achieve stress
effects through combinations. The means to combine fall into the following: 1) Os.Polar +
Os.Linear indicates near-synonym. For example, fan in polar coordinates form a pie diagram
strip rectangular linear, and then form into a stacked column symbol. Pie charts can highlight
structure attributes, columnar symbol can highlight quantative attributes, and these two can
stress thematic elements; 2) Os.Polar + Os.Polar refers to synonym and near-synonym. For
example, a fan-shaped construct of two semi-circles in the clockwise direction and counter
clock wise direction. The change of shape or color, or different numerical weights can
emphasize different elements; 3) Os.Polar + Os .Topological refers to synonyms and
near-synonyms. For example, a fan-shaped polar coordinate construction expresses the
construction of thematic elements, and geometric polygon of thematic elements through
topological construction and by adopting the classified method, constructs classification
maps to indicate contrasting quantitative spatial distribution.
Figure 5a indicates statistical information on basic education in an administrative area. It is a
combination of two semi-circles from top to bottom. The top semi-circle indicates quantity
and composition of basic education agencies in this administrative area, the bottom
semi-circle indicates student enrollment and composition of basic education agencies in this
administrative area. The meanings of the synonym combination is merely similar but not
same. In these instances, the number of schools and students is relative. In general, if there
are more schools, the more students. So using semi-circle to contrast,the radius ratio may
have an effect on cognition.
2. Hyponymy combination:it includes hierarchical combination and same class combination.
The former one is a combination of superordinate and hyponym, it indicates the degree of
hyponym which includes its superordinate. For instance, a combined expression of
agriculture and farming indicates the condition of farming in agriculture. The latter is a
hyponym combination indicating superordination in the same class. For example, a
combined expression of farming and animal husbandry refers to agriculture development at
the same level. Two kinds of combination uses polar coordinate construction. Because
quantity units between words may be not unanimous, the coordinate axis for linear
construction is unique whether or not it is a vertical combination or horizontal combination.
The
combination
methods
are
as
:
Os.Polar(…,Polar.startAngle,…)
+
Os.Polar(…,Polar.startAngle + spj. Angle*p,… )，and p∈(0,1).
Figure 5b shows atmospheric condition over some years in a city, it has three indexes;
sulfur dioxide, nitrogen oxides and particulate matter. Three indexes as a hyponym vertical
combination convey the hyponymy relations of superordinate atmospheric conditions. Two
combinations are included. The three indexes for one year are generated by the polar
coordinate construction ofa fan-shaped morpheme, different sizes of scales indicate
different quantative attributes of quantity units and similar color zones indicate differing
hyponymy; when structuring polar coordinates for different years, a changed start angle is
combined with the last year’s and brightness to indicate time attributes.
3. Meronymy combination:it can show part-whole relation of thematic contents or between
contents. It includes a whole data feature and part of a structural feature, reflecting a whole
quantative feature through linear construction (or fixed construction). A polar coordinate
describes the partial structure of thefeature. The two will be generated by fixed construction.
Its basic combination formation is: Os.Point(Os. Polar, Os. Linear). In Figure 5c, a columnar
symbol refering to total population is generated by fixed construction, and a ring symbol
refers to the composition of the total population. In this combination, the dimension
variable is used in “whole word” to reflect whole numerical content of thematic elements,
and the “part word” adopts a color variable to distinguish composition features; a its size
does not have numerical meaning.
4. Antonymy combination: refers to the contrary relation of objects. It can achieve a
prominent effect or counterbalancing effect through contrast. Antonymy relations reflects
binary antonymy (such as male and female in sex composition) and relation antonymy
(financial revenue and expenditure). An antonymy combination adopts a linear construction.
When conveying binary antonymy , the data direction of words and construction should meet
at right angles for a bipolar orientation such as a pyramid chart. When reflecting relation
antonymy, the direction of data must be the same, such as in linear statistical chart and
planar charts. Its basic combination is: Os.Point (Os.Linear, Os.Linear).
In thematic maps, a pyramid symbol that reflects population composition of age and sex
adopts a linear construction to describe different age groups of the male population and
female population. Its direction of data is opposite to the construction and is constructed by
fixed construction to manifest contrast between different groups of males and females. In
Figure5d, a combination of different plane domain word symbols reflects financial revenue
and expenditures over recent years. This combination can highlight financial conditions
easily, purple represents the local general budget revenue, green represents the general
budget expenditure. From the figure, revenue is more than expenditures, and increases every
year.
a)
c)
b)
d)
Figure 5. Combination principles in sentence level
4.
Automatic Construction Mechanism of Thematic Symbol
Currently, most thematic symbols on desktop or web mapping systems follow the guided or
templet mapping mode. This provides limited default thematic symbols and then conducts
mapping through parameterized settings. This method reduces mapping difficulties and
increases the number of users. Based on the symbolic construction model presented
previously, this section will focus on the design of universal and extendible mapping models,
including formalized description and automatic construction mechanisms.
4.1. Formalized Characterization of Syntax Structure
Design mode conduct representation aiming at the foundational mechanisms for automatic
symbol construction. When using in online thematic mapping, formalized characterization
should be conducted to share service models (Figure 6). TSentence is root node of a model
structure diagram, and its namespace is Thematic Symbology Encoding,tse. Corresponding
to the sentence, word and phoneme, it contains construction principle, semantic relations
between words, tse:TCartoWord and tse:TCartoPhoneme. TCartoPhoneme contains two
sub-elements: Primitive Type and Construction Property. Idiomatic morphemes can form
words directly which involved in construction at the syntax level, and non-morphemes can
construct into words through the construction principles and visual variables. TCartoWord
contains three sub-elements which are mapping phonemes, construction principles and
visual variables, mapping sentence set can be combined through semantics between words.
Fig.6 The XML schema of thematic symbol based automatic construction mechanism
Four symbolic patterns (TPointSymbolizer、TLineSymbolizer、TPolygonSymbolizer and
TChartSymbolizer) were integrated through syntax construction. Symbolic elements such as
point, line and polygon follow the Symbol Encoding (SE). The points is a fixed proportional
symbol. The line is made from linear primitives or polyline primitives constructed through
topology, and defined primitive pattern and construction methods. TPolygonSymbolizer can
be divided into TcartogramSymbolizer and TChoroplethSymbolizer, corresponding to the
topological structure diagram and choroplethic map respectively in the topology construction.
Among them, TCartogramSymbolizer(Figure 7) contains three sub-elements that are
CartogramType、PropertyName and DistortionQuality. CartogramType is used to describe
algorithm type of a topological structure diagram such as Diffusion(Gastner & Newman,
2004), CartoDraw(Keim et al., 2004), Carto-SOM (Henriques et al., 2009) and etc., and the
sub-element CartogramTypeName decides its value through enumeration. PropertyName
can define the ID of layers which are required to build, and it is numeric type.
DistortionQuality describe the quality characteristics generated by a topology construction
algorithm in the topological structure diagram, its value range is from 1 to 100. Additionally,
the value range will influence algorithm precision. A high value means much more iteration
time for topological transformation and a better shape quality . But it may cost more time
when implementing the algorithm. Low value means less iteration time, but the shape may be
distorted.
Fig.7 The XML schema of cartogram symbol
Statistical charts can be divided into chart type (ChartTypeName) as previously defined and
symbol type (TSentence) constructed automatically by the symbolic construction mechanism.
ChartTypeName presents the corresponding symbol name and string type. It can be selected
from the predefined appointed symbol names and cannot be present more than twice. So
symbolism support combines solely for two kinds of symbols. TSentence is symbol type for
automatic construction. The symbols generated by this method can record construction
parameters, add predefined types, and provide calls for other users.
4.2. Automatic Construction Mechanism for Thematic Symbols
This section focuses on analyzing the thematic semanteme property, confirming syntax
construction of syntax symbols and constructing automatic construction mechanism by
constructing a thematic indicator. Based on current data and ontology models(Andrienko
and Andrienko, 1999;ZHAO et al., 2011), Zeng et al., 2013), thematic indicators in
heterogeneous systems can be define as follows: Thematic Indicator Ontology = {S, T, C1, C2,…,
Cn, I, U, Corr}. In this formula, S(Space) represents spatial scale; T(Time) represents time
scale; C(Category) represents property classification; I(Indicator) represents thematic index;
U(Unit) represents unit scale; Corr(Correspondence) represents semantic relations that can
be abstracted: corr:P→R∪I∪{(Ii, {(rj, vj,k)}, 1≤ j ≤ m, vj,k ∈Rj), 1≤ i≤n}. R(Refence)
indicates reference property including indicator properties (S、T、C，I). (Ii,{(rj, vj,k)})
represents a certain specific thematic indicator, and (rj, vj,k) represents the reference property
set of the constraint index Ii.
DISTRICT CITY YEAR POPULATION GENDER ZERO_FOUR ……
……
…… …… ……
……
……
……
EIGHT_ABOVE
……
Fig.9 The illustration of Thematic Indicator Ontology Model
Figure 4 contains five reference attributes and two index attributes. Thereinto, r1、r2 are
space attribute, r3 is time attribute, r4、r5 are category attribute. I1 is aggregate indicator
which is not limited by category attribute, I2 is structure indicator and its value is constrained
by two category attributes. For I2, the steps of automatic construction mechanism for
word-centered shown as: ① Quantity of elements for each level of syntax can be confirmed
according to reference attribute. The spatial reference attribute decides the number of
sentence. Category reference attribute decides the number of words when constructing each
sentence (If there is no category reference, the default is 1). Time scale and indicator
dimension decides the number of phoneme of each word (i*j, i is the number of attribute
categories，j is the number of attribute variables). ②A category attribute of indicators can
define construction levels and construction principles, in this example, r4 in I2 is dominant
restraint and the semanteme is autonymy, r5 is implicit constraint, and the semanteme is
synonym. Therefore, word level employs linear or polar coordinates construction according
to year, and the syntax employs the linear inverse construction according to gender. ③ In
order to provide interactive selection, the system filters thematic primitive types that are
suitable for word-level construction based on the primitives construction property. In this
case, linear construction employs columnar primitives. ④ Combine with visual variables to
generate complete words. Then, the words are added as annotation or text tips. Words with
different genders generate combination symbols according to the combination principles of
the syntax dimension. ⑤ Finally, complete symbols can be drawn on the map according to
given location coordinates.
The whole process for automatic construction of thematic symbol is shown in Figure 8.
Firstly, thematic element is population composition of certain city in 2005, and its indicator
ontology analysis shows in Figure 4. Selecting indicator I2, choosing a thematic
primitives—rectangle, and then combining with a visual variable—colors form two phonemes
Sp1、Sp1; Secondly, according to liner construction structure principle Os at the word level
mentioned previously, adjusting the interval of Sp1 to gain the best style of symbols; Defining
Sp and Os, then generating words Sw1、Sw2 automatically. This syntax level employs binary
antonymy construction, and realizes through the rotation property in positioning
construction. Users can control the parameters, and further construct Ss according to labels
and text tips of Sp. In this processing, background symbols generation employs software
system in 4.1, the foregrounding can be formalized into symbolic characterization documents
which can be provided to different platform. In this case, users can select word combination
modes at different stages. For example, phoneme rotation parameter in controlled linear
construction set as 90°and -90°respectively, and then syntax level can place a word directly
followed by positioning construction.
Fig.10 The automatic construction of thematic symbol based on the syntactic theory (triangle chart for example)
5.
Conclusion
The proposed theory provides a new innovative idea for symbol construction. It has been
successfully applied in China for Online Dynamic Atlas of Shenzhen (ZHAO et al., 2011)
and Zhejiang Online dynamic thematic map production software (ZENG et al., 2012).
Furthermore, online interactive mapping model and web thematic cartography services have
theoretical and realistic reference meaning to the automatic cartography for distributed data.
In this mechanism, symbols can be defined based on three levels which are letter, word and
sentence. Each level can convey dynamically and intelligently the essential thematic factors
for many dimensions through integration of primitive construction principles and retinal
variables. furthering future research, automatic recognition and intelligent selection of
multiple semantic spatial information, and construction mode deserves more attention. The
recognition effects of construction symbols based on this mechanism and measurement of
amount of information are also significant areas to be explored.
References
ANDRIENKO G, ANDRIENKO N 1999. Data characterization schema for intelligent
support in visual data analysis [M], Spatial Information Theory. Cognitive and
Computational Foundations of Geographic Information Science. Springer:
349-365.
ARMSTRONG M P, XIAO N, BENNETT D A 2003. Using genetic algorithms to create
multicriteria class intervals for choropleth maps. Annals of the Association of
American Geographers [J], 93(3): 595-623.
BERTIN J 1983. Semiology of graphics: diagrams, networks, maps [M]. University of
Wisconsin Press; Madison.
BOULOS M N K, WARREN J, GONG J, YUE P. 2010. Web GIS in practice VIII:
HTML5 and the canvas element for interactive online mapping. International
Journal of Health Geographics [J], 9(1): 14.
DU Q 2004. The Microlinguistic Conceptual Model of Spatial Information.
GEOMATICS WORLD [J], 2(6): 5-20.
FERN NDEZ P. Paradigmatic Tendencies in Cartography: A Synthesis of the
Scientific-Empirical, Critical and Post-Representational Perspectives [D].
Dresden; Dresden University of Technology, 2012.
GASTNER M T, NEWMAN M E J 2004. Diffusion-based method for producing
density-equalizing maps. Proceedings of the National Academy of Sciences of
the United States of America [J], 101(20): 7499-7504.
GUO D 2009. Flow mapping and multivariate visualization of large spatial
interaction data. Visualization and Computer Graphics, IEEE Transactions on
[J], 15(6): 1041-1048.
HENRIQUES R, BACAO F, LOBO V 2009. Carto-SOM: cartogram creation using
self-organizing maps. International Journal of Geographical Information
Science [J], 23(4): 483-511.
IOSIFESCU-ENESCU I, HUGENTOBLER M, HURNI L 2010. Web cartography with
open standards - A solution to cartographic challenges of environmental
management. Environmental Modelling & Software [J], 25(9): 988-999.
JANKOWSKI P 2009. Towards participatory geographic information systems for
community-based environmental decision making. Journal of Environmental
Management [J], 90(6): 1966-1971.
JIANG X 1986. THE POSITIONAL SYMBOL SYSTEM OF COMPUTER-ASSISTED
THEMATIC MAPPING AND AUTOMATIC DRAWING OF DERIVED
SYMBOLS. Acta Geodaetica et Cartographica Sinica [J], 15(3): 196-206.
KEIM D A, NORTH S C, PANSE C 2004. CartoDraw: A fast algorithm for generating
contiguous cartograms. Ieee Transactions on Visualization and Computer
Graphics [J], 10(1): 95-110.
LI H, WANG Y, YU Z, HAN J, LUO B 2009. Applying Case-based Reasoning
Techniques in Selection of Representation Method of Statistic Map.
JOURNAL OF GEO-INFORMATION SCIENCE [J], 11(6): 819-825.
LI W, ZHOU Y 1997. The Design and Implementation of a Symbol Base in Thematic
Mapping. Journal of Wuhan Technical University of Surveying and Mapping
[J], 22(3): 263-265.
MA Y 1995. A STUDY ON THE SYMBOLIC CONSTITUTION ELEMENTS AND ITS
DESIGNING MODELS. Acta Geodaetica et Cartographica Sinica [J], 24(4):
309-315. HUA Y 1993. Determine Map Content Of Thematic Map With
Expert System Technology. Journal of Institute of Surveying and Mapping [J],
20(1): 56-61.
SCHNABEL O, HURNI L 2009. Primitive-based Construction Theory for Diagrams
in Thematic Maps. Cartographic Journal [J], 46(2): 136-145.
SHIMIZU E, INOUE R 2009. A new algorithm for distance cartogram construction.
International Journal of Geographical Information Science [J], 23(11):
1453-1470.
TIAN J, HUANG R, GUO Q 2007. Study on intelligent choice of representation
methods in thematic map. Science of Surveying and Mapping [J], 32(5):
170-172.
WILKINSON L, WILLS D, ROPE D, Norton A, Dubbs R 2006. The grammar of
graphics [M]. Springer.
YU L 1995. The Philosophical Levels of Map Symbol and The Exploratidn of Its
Information Function. Acta Geodaetica et Cartographica Sinica [J], 24(4):
259-266.
ZENG W, ZHAO F, REN F, DU Q. 2012. Thematic Cartography Based on JFreeChart
and ArcGIS JavaScript API. Journal of Geomatics Science and Technology [J],
29(6): 450-453.
ZENG X, DU Q, REN F, ZHAO F. 2013. Design and implementation of a web
interactive thematic cartography method based on a web service chain.
Boletim de Ciências Geodésicas [J], 19(2): 172-190.
ZHAO F, DU Q, PENG Z, REN F. 2011. Interact ive Model for Web Thematic
Cartography: A Indicator-driven and Task Flow-centered Approach. Acta
Geodaetica et Cartographica Sinica [J], 40(5): 655-661.
ZHU D 1988. On Map Structure. Acta Geodaetica et Cartographica Sinica [J], 17(2):
151-158.