POI Symbol Design in Web Cartography—A Comparative Study

Abstract

This paper studies the design of point symbols on widely used online maps and apps that portray tourist points of interest (POIs). Tourist maps are among the most commonly used types of maps nowadays. The ease of travel leads to an ever-increasing demand for tourist maps. Therefore, appropriate map design, content and technical means are necessary for better information transfer and communication between the map and the user. Online maps and apps were selected according to specific criteria (e.g., language, geographic area, pictographic symbols, interactivity). Pictographic point symbols for POIs related to tourism activities were collected and described by variables. The frame and the pictogram of the point symbols were the two main directions for choosing the descriptive variables. Description is based on the cognitive scheme for interpretation of cartographic symbols with some improvements. The study reveals the characteristics of the point symbols and constitutes a documentation of the applied practices. The main trends prevalent in the design of these symbols are highlighted and commented on in relation to traditional cartographic practices and guidelines for the design of point symbols.

1. Introduction

Nowadays, tourist maps are an integral part of travels all over the world. They offer geographical and tourist information and are essential for tourists exploring relevant destinations [1]. They constitute a special type of map and the most complex area to work with in cartographic concepts and processes [2]. Special attention was paid to tourist maps in the 1970s and 1980s when scientific conferences addressed symbol standardization, map legends, content representation and interpretation, maps in travel guides, cartography contribution to tourism development, etc. [3]. In recent years, their use seems to have increased, as the emergence of the Internet, combined with easier access to short trips and unknown destinations, has encouraged the use of web maps and map applications. People’s lifestyle today favors short trips and the need to visit unfamiliar destinations. As the demand increases, optimizing the design of tourist maps becomes an important issue of map effectiveness.

The main challenge in designing a tourist map is that it must serve different purposes and users [2]. With an effective map design, tourist maps can encourage tourists to experience more than just the key urban landmark attractions [4]. There is a link between cartography and tourism, a scientific discipline and a practical field, as both are concerned with the development, production and use of maps [5]. A well-designed map provides tourists with explicit orientation and is easy to use; a poor design, on the other hand, might waste users’ time and even mislead them [6]. Nowadays, a number of maps suffer from poor design [7]. Travel map design deserves scientific attention for several reasons: scale, appropriateness of cartographic symbolism and text, the effectiveness of graphic language, layout, the inclusion of insets and photographs, depiction of transportation and tourism infrastructure and illustration of parks and recreation areas [8]. Symbol design is crucial for reading tourist maps. Point symbols are particularly significant on tourist maps because they usually carry the thematic content, indicating points of interest (POIs) that should attract the attention of the user; thus, the adoption of an original graphic design can make them attractive for the users [9].

Point symbol design is discussed at length in the cartographic literature [10]. Forrest and Castner (1985) [11] designed and performed an experiment on the perception of point symbols for tourist maps for the first time and concluded that pictorial symbols are the most effective, and framed symbols are visually clearer and held out when a background color is applied as well. Clarke (1989) [12] argues that cartographic symbols are effective if they are easily and accurately understood by users and favors symbols that are already in usage, introducing the aspect of symbol familiarity. In addition, the degree of internal complexity is linked with symbol size [12,13]. Appropriate size may shorten the time required to find a pictorial symbols on a touch screen in the context of tourist maps [14], although rapidity is also impacted by familiarity [13]. Finally, graphic content clarity and simplicity are also important [15,16]. All point symbols for a map should be designed as an entire set to have consistent visual complexity [17]. In addition to symbol design, symbol interpretation is influenced by map content and nearby symbols [15]. Finally, the link of symbol design with standards is raised by Gerber et al. (1990) [15].

From another perspective, point symbols must be straightforward to facilitate understanding of their meanings [18]. Gerber et al. (1990) [15] introduce the resemblance to the referent. Clarke (2013) [12] investigated through questionnaires on two comparable published tourist maps and concluded that point symbols, i.e., abstract, geometric or pictorial symbols, can be inefficient if they bear a poor resemblance to the object they depict. This statement is in accordance with i. Ratajski’s (1973) [10,19] principle of forms isomorphism, e.g., shapes are similar to objects or the objects attributes; ii. MacEachren’s [20] suggestion that the best recognizable cartographic symbols directly represent a specific object or refer to associations relating to it and iii. the imitation-based design described by Keates (1996) [21]. Pictographic symbols can be categorized further on the basis of how the symbol (i.e., the visual representation) is linked to its meaning (i.e., the referent) [22]. Nakamura and Zeng-Treitler (2012) [23] provided a taxonomy of the representation strategies of pictographic symbols: arbitrary convention (e.g., a cross for a hospital), visual similarity (e.g., a building with a cross for a church) and semantic association (e.g., cutlery for a restaurant).

From the above, one can understand that the success of the pictorial symbols is related to pictogram (icon) selection. Icon characteristics include concreteness, complexity, meaningfulness, familiarity, semantic distance, distinctiveness and aesthetic appeal [24,25,26]. The abovementioned icon characteristics can be grouped as graphic (simplicity, clarity, visibility, consistency, distinctiveness, aesthetic appeal) and semantic (concreteness, semantic closeness, familiarity and acceptability) [22]. These qualities are documented in cartographic but also in the behavioral literature through research [27]. However, the assessment of the applied basic geometric and graphic variables, visibility, concreteness and symbol dimensions can be accurately measured, but other conditions, such as the assessment of simplicity, familiarity and semantic closeness of cartographic symbols, depend on the subjective impression, background and abilities of the observer [27].

Pictographic symbols are considered the best option for tourist maps [11,13,28] since they are more eye-catching than traditional topographic symbols [29]. Moreover, image-related symbols should be used whenever possible because their comprehension is less dependent on the cultural context [30]. They can be identified without a legend, but the symbols must be correctly interpreted before claiming this benefit [31]. Some researchers have proposed practical guidelines for pictorial symbol design, such as using a frame and a background [11]. Notable design decisions in GeoVista mobile map symbols include [17]: a square frame that provides space to increase the icon size and therefore legibility, a light icon within a dark frame that provides a high figure–ground contrast with base maps which are typically light in color, a strong semantic relationship with objects being depicted and a slight shadow that promotes figure–ground relationships and helps to separate the symbols from the base map with a raised appearance. In another study [9], guidelines to ensure consistency include a frame of consistent width and signs as filled figures that occupy a similar area and have geometric or irregular forms. More precisely, a square black frame of 0.2 mm is proposed, a 5–7 mm diameter and a black sign on a yellow background.

This paper aims to study the point symbols on web maps and applications that portray tourist information. Symbol description is based on the cognitive scheme for interpreting cartographic symbols by Bertin [32] and MacEachren [20]. The main trends prevalent in the design of these symbols are highlighted and commented on in relation to traditional cartographic practices and guidelines for the design of point symbols.

The rest of the paper is organized as follows. Section 2 describes the criteria for the selection of online tourist and general maps and the methodology for describing the point symbols. Section 3 presents the symbols of tourism POIs collected from the maps and symbol description based on the proposed variables. Section 4 presents the results of the data analysis at global, map and POI level. Finally, results are commented on and conclusions are presented in Section 5. In Appendix A, POI symbols are presented, and in Appendix B the calculated percentages of the variables for the symbol description are given.

2. Methodology

2.1. Data Acquisition

The aim of this study was to investigate point symbols for tourist information on online maps and apps that use the English language and portray the European continent. The first step was to select a number of maps on which the study was based. The criteria adopted for the map selection were:

• Pictographic symbols: Maps should include point symbols that portray tourist information, e.g., a restaurant, an airport, a castle or simply a sight view. This research focuses only on pictographic symbols; consequently, maps that symbolize the points with geometric symbols (e.g., circles) were not included. Pictographic symbols constitute the best choice for tourist maps, as they produce less uncertainty than geometric symbols, and their meaning can be intuitively grasped by those who are inexperienced in reading maps and even without a legend.
• Type of access: Nowadays, tourists may use maps in websites or in applications.
• Free and easy access: Easy-to-access web maps and free apps tend to be the ones most used by tourists.
• Popularity: The more frequently users use a map, the more that map seems to meet their needs. Unique visitors per month can measure popularity. Popularity data were collected from “Similar Web” [33] for March 2023. Popularity is an important criterion but not a reason for excluding a map from the survey.
• High cartographic quality: Some internet maps are characterized as “low quality” as they do not meet esthetical and cognitive aspects as well as aspects of communication assured by graphical design [34]. Cartographic quality is also supported by symbol qualities such as visibility, simplicity and clarity of the symbols [22].
• Interactivity: Maps should have basic navigation tools that appear in slippy maps such as “pan” and “zoom”. Additionally, information on the geographic entity portrayed by the point symbol should be provided especially when the legend is missing (e.g., Google Maps).
• Map Type: Tourist maps on the web are provided by Travel Companies and National Tourism Organizations. However, general-purpose online maps (e.g., Bing Maps) also have a leading position in travel planning [35].

After a survey in the web for maps and apps that can be used by tourists, 23 maps were assessed (Google Maps, Bing Maps, MapQuest, Open Street Map, Map.Naver, 2gis.ru, Map.baidu.com, Maps.me, Sygic Travel Maps, Offline Map and Travel Guide, Roadtrippers, Visit a City, Use-It Travel, Orange Smile, Here We Go, Search.ch, Mytouristmaps.com, Visit Norway, Visit Estonia, Esmadrid, Germany.travel, France.fr, Welcome to Warsaw (Missouri) (accessed on 30 March 2023)). Based on the above criteria, 13 were selected. Considering that the study area is Europe, “MapQuest”, “Welcome to Warsaw” and “Map.Naver.com” were rejected since they do not cover this area. The “2gis.ru” map is not provided in the English language, as is also the case for “Map.baidu.com”. “Visit a City” and “Germany.travel” were rejected because they do not use pictographic symbols but geometric. Finally, “Use-It Travel”, “Mytouristmaps.com” and “EsMadrid.com” do not have pan and zoom functions. The selected maps (

Table 1. Selected maps.

Map	Technology	Provider	Type
Google Maps (GM)	Websites	Map companies or VGI initiatives	General
Open Street Map (OSM)
Bing Maps (BM)
Maps.me (MM)	Applications	Companies for travel and tourism	Tourist
Sygic Travel Maps (STM)
Offline Map and Travel Guide (OMT)
Roadtrippers (RT)
Orange Smile (OS)	Websites	Companies for travel and tourism	Tourist
Here We Go (HWG)
Search.ch (SCH)
Visit Norway (VN) Visit Estonia (VE) France.fr (FFR)	Websites	National Tourism Organizations	Tourist

) are both general maps by map companies or VGI initiatives and tourist maps provided by National Tourism Organizations or commercial tourist companies. A number of them are applications and other websites. They constitute a representative sample of maps providing tourist information on the web.

After the map selection, maps were reviewed to find the most common POIs related to tourism that are portrayed. A number of tourist POIs were assessed for the study (35). At a second stage, all POIs were classified in relation to their thematic content, e.g., “hotel”, “motel” and “hostel” are three different POIs which belong to the category “accommodation”. Therefore, some general content categories, e.g., “accommodation” are created to classify the tourist POIs (e.g., hotel, hostel, motel). The criteria to select and classify the tourist POIs are the available material in map sources, the content of today’s tourist maps and existing studies [13,36]. Finally, the 35 tourist POIs were grouped into eight categories (

Table 2. Categories and tourist POIs.

Category	Tourist POI
Leisure–Recreation	Sport activities Water park Sights Zoo
Accommodation	5. Camping 6. Guest accommodation 7. Hostel 8. Hotel 9. Motel
Culture	10. Arts-center gallery 11. Cinema 12. Theater 13. Library 14. Museum
Gastronomy–Entertainment	15. Café 16. Fast food 17. Bar 18. Pub 19. Restaurant
Shops	20. Clothing store 21. Department store 22. Gift shop 23. Grocery store 24. Shoe store 25. Shopping mall 26. Supermarket
History–Religion	27. Castle 28. Church 29. Monument
Transportation	30. Airport 31. Bus stop 32. Ferry service 33. Gas station 34. Train station
Healthcare	35. Hospital

).

2.2. Methodology for Symbol Analysis

The study of symbols used in maps to portray POIs must be based on a description in cartographic terms, e.g., visual variables, and in a systematic way that will permit data analysis. The term visual variable is commonly used to describe the various perceived differences in map symbols that are used to represent geographic phenomena [37]. The concept of visual variables was introduced by Bertin [33] and subsequently modified by others [21,38,39]. Bertin [32] considered the viewing of these visual variables as a pre-attentive process [40]. They create a mental image or communicate meaning before the map viewer’s internal image schemata are brought into play [41]. Visual variables also play an essential role in the design of cartographic symbols for augmented reality (AR) applications on smartphone-type mobile devices [42] and video games [43]. Our work considers visual variables for static maps, i.e., size, shape, orientation, color hue, color value and pattern.

MacEachren [20] proposed a cognitive schema for interpreting point symbols on the National Park Service map. This schema can be used as a framework for point symbol analysis. The symbol is divided into the “frame” indicating the position and the “interior” related to the activity or facility. Both elements can be described by visual variables such as color, shape and size. This idea is further elaborated [44] for the interpretation of symbols on crisis maps: the “interior” is replaced by the “pictogram”, and an outline is added to the frame description. While adapting this schema to the symbol description in this study, the list of visual variables was enriched with basic hue, frame pinned and frame background. The final variable list [45] consists of the basic hue and symbol size, elements for frame description, i.e., frame outline, frame outline color, frame shape, frame pinned, frame background, frame background color and elements for pictogram description, i.e., pictogram shape and pictogram color.

Color is the most important visual variable for expressing spatial information in cartographic visualizations. Proper use of color can clearly highlight the content of a map and improve its readability and comprehensibility. For these reasons, basic hue was introduced as a symbol’s dominant color, capturing the color’s visual impact to the user. The size of a symbol refers to the horizontal dimension and expresses the desktop or the mobile size depending on the map source. Frame outline, the third variable, distinguishes the symbol from the map background, and it is therefore a critical design factor. A closed form of the area of interest is one of the graphic design principles for a successful figure–ground organization [46]. The frame outline color refers to the color, and the frame shape refers to the shape, which is usually a simple and compact form such as a circle or a rectangle. The frame outline color and the frame shape can communicate higher-level categories of point symbols [47]. The frame pinned variable is related to the existence or not of an anchor point in the frame outline (e.g., the circle is not a frame-pinned scheme). The frame background refers to the existence or not of a background behind the pictogram and the frame background color to the color. The frame background color is essential since it can attract the user’s attention. A high contrast between the pictogram and the background color (e.g., a black symbol on a white background) facilitates search and avoids confusion between the figure and its background [48], while a low contrast (e.g., a black symbol on a dark gray background) leads to poorer results. The pictogram color is a prominent dimension in defining targets for search, being more proficient than shape, size or alphanumeric characters in defining targets [49]. The pictogram shape refers to the form of the icon. It is considered the most common visual variable that map icons use to differentiate between referents [50]. It can be regarded as pictorial or mimetic, abstract or associative. Based on the collected symbols, pictograms are classified into abstract design, e.g., star with border; alphabet letters, e.g., H for hospital; and pictorial icons. Pictorial icons are anthropocentric and refer to items used in daily life, buildings, human activities or the natural environment. They have a great degree of concreteness, which refers to the extent to which the symbol depicts real objects, materials or people [24]. Additionally, semantic closeness, that is the measure of the distance between the map symbol and what it is intended to represent [22,24], or iconicity, that is the degree to which a sign vehicle represents its referent [32], is highly desirable. Pictogram shapes are semantically classified into six groups, i.e., items, edible items, transportation, buildings, activities and nature. Finally, pictogram color refers to the color of the pictogram. The selected variables, the values and a short explanation are presented in

Table 3. Point symbol description: variables, values and a short explanation.

Variable	Values	Explanation
Basic Hue	Black (BK), blue (BL), brown (BR), green (GR), gray (GY), orange (OR), pink (PK), purple (PR), red (RD), yellow (YL)	The dominant color of the symbol
Size	2–9 mm	Desktop or mobile size depending on map source
Frame Outline	Yes (1)/no (0)	The closed contour of the symbol (exists or not)
Frame Outline Color	Black (BK), blue (BL), brown (BR), green (GR), Gray (GY), orange (OR), pink (PK), purple (PR), red (RD), yellow (YL)	The color of the closed contour
Frame Shape	Pin, circle, square, speech bubble	The shape of the closed contour
Frame Pinned	Yes (1)/no (0)	Whether the closed contour has an anchor point or not
Frame Background	Yes (1)/no (0)	Whether a background exists or not behind the pictogram
Frame Background Color	Black (BK), blue (BL), brown (BR), green (GR), gray (GY), orange (OR), pink (PK), purple (PR), red (RD), yellow (YL)	The color of the background
Pictogram Shape	Abstract design (ABST), alphabet letters (LETR), nature (NTR), building (BUILD), item (ITEM), activities (ACTIV), transportation (TRSP), edible items (EDBL)	The form of the icon that exists within the frame
Pictogram Color	Black (BK), blue (BL), brown (BR), green (GR), gray (GY), orange (OR), pink (PK), purple (PR), red (RD), yellow (YL)	The color of the icon

. An example of a point symbol description based on the above variables can be found in Figure 1.

Based on this methodology, the point symbols can be described objectively and comprehensively. Most of the variables are in nominal scale, and are coded with alphanumeric characters, i.e., “yes”/“no”, color names and specific content types (buildings, animals, etc.). Only the size variable holds numerical values (ratio scale). Data analysis of symbols across maps is based on calculating frequencies and percentages for the variables.

3. Material

Each map contains many point pictographic symbols that differ in style and design characteristics (Figure 2). A complete list of the 390 collected symbols can be found in Figure A1 and Figure A2 in Appendix A. Some of the selected tourist POIs may not be portrayed in all maps (

Table 4. Tourist POIs and Maps.

Map	Percentage of Tourist POIs	Number of Tourist POIs
Google Maps (GM)	97%	34
Open Street Map (OSM)	91%	32
Bing Maps (BM)	97%	34
Maps.me (MM)	94%	33
Sygic Travel Maps (STM)	100%	35
Offline Map and Travel Guide (OMT)	83%	28
Roadtrippers (RT)	69%	26
Orange Smile (OS)	83%	29
Visit Norway (VN)	89%	31
Visit Estonia (VE)	89%	31
Here We Go (HWG)	100%	35
France.fr (FFR)	66%	22
Search.ch (SCH)	60%	20

). Most maps include at least 80% of the studied POIs, and three have a substantial percentage of 60%. Results can be considered representative of the symbol design for the studied maps due to these percentages. Based on the adopted method for symbol description, all symbols can be described with the selected variables and were analyzed in the subsequent stages. For example, in

Table 5. Description of the “Camping” symbol for all maps.

						Maps
	GM	OSM	BM	MM	STM	RD	VN	VE	HWG	SCH
Variable
Basic Hue	GR	BL	GR	GY	GR	GR	BK	GR	BL	PK
Desktop Size (cm)	9	4	5.5	6	5	3	4	6	10	4
Mobile Size (cm)	5	3	4	3	3	3	3	4	9	1
Frame Outline	Yes	No	Yes	Yes	No	Yes	No	No	Yes	No
Frame Outline Color	GR	-	GR	WH		WH	-	-	PR	-
Frame Shape	Pin	-	Circle	Circle	Circle	Circle	-	Circle	Pin	Circle
Frame Pinned	Yes	-	No	No	No	No	-	No	Yes	No
Frame Background	Yes	No	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes
Frame Background Color	GR	-	GR	BR	GR	GR	-	BL	PR	PR
Pictogram Shape	Item	Item	Item	Item	Item	Item	Trsp	Item	Item	Item
Pictogram Color	WH	BL	WH	WH	WH	WH	BK	WH	WH	BK

, one can see the description of the “Camping” symbol for a number of maps participating in the case study.

4. Data Analysis—Results

Analysis of the variable values that describe the symbols was performed with the statistical software Jamovi v.2.3.28 [51]. Initially, the general analysis examined each variable (e.g., basic hue, size) for all maps to investigate the point symbol design. Finally, a special analysis was applied for the use of pictograms.

4.1. General Data Analysis

The first stage of the data analysis examined the behavior of each variable for all symbols (390) in all the maps. The most popular basic hue is blue (26.4%), and the least popular is yellow (1.3%) (Figure 3,

Table A1. Frequencies and percentages (%) of basic hue.

	Frequency	Percentage (%)
Black (BL)	33	8.5
Blue (BL)	103	26.4
Brown (BR)	53	13.6
Green (GR)	57	14.6
Gray (GY)	12	3.1
Orange (OR)	44	11.3
Pink (PK)	26	6.7
Purple (PR)	42	10.8
Red (RD)	15	3.8
Yellow (YL)	5	1.3

).

The majority of symbols (63%) use a frame outline (

Table A2. Frequencies and percentages (%) of frame outline.

	Frequency	Percentage (%)
YES	245	62.8
NO	145	37.2

). The most popular frame outline colors are white (24.1%) and green (17.1%) (Figure 4,

Table A3. Frequencies and percentages (%) of frame outline color.

	Frequency	Percentage (%)
Black (BL)	30	12.2
Blue (BL)	37	15.1
Brown (BR)	10	4.1
Green (GR)	42	17.1
Gray (GY)	2	0.8
Orange (OR)	11	4.5
Pink (PK)	19	7.8
Purple (PR)	13	4.9
Red (RD)	14	5.7
White (WH)	59	24.1

). Gray is the least used color (0.8%). White can be easily distinguished from the map background and is considered an effective boundary color even when overlaps exist [52]. The circle is the most common frame shape (65%) (Figure 5,

Table A4. Frequencies and percentages (%) of frame shape.

	Frequency	Percentage (%)
Speech bubble	29	9.6
Circle	196	65.1
Pin	64	21.3
Square	12	4

). Most symbols (62.1%) do not have a pinned frame (

Table A5. And percentages (%) of frame pinned.

	Frequency	Percentage (%)
YES	114	37.9
NO	187	62.1

).

The vast majority (77.2%) have a frame background (

Table A6. Frequencies and percentages (%) of frame background.

	Frequency	Percentage (%)
YES	301	77.2
NO	89	22.8

), a clear way to separate the symbol from the map background. The existence of a frame background is unrelated to the presence of a border, although the frame background is schematically limited. Regarding the background color, only the symbols with a background are evaluated. The most common background color is blue (30.6%) (Figure 6,

Table A7. Frequencies and percentages (%) of frame background color.

	Frequency	Percentage (%)
Blue (BL)	92	30.6
Brown (BR)	22	7.3
Green (GR)	46	15.3
Gray (GY)	12	4
Orange (OR)	34	11.3
Pink (PK)	26	8.6
Purple (PR)	31	10.3
Red (RD)	14	4.7
White (WH)	19	6.3
Yellow (YL)	5	1.7

), the second is green (15.3%) and the least common is yellow (1.7%).

The pictogram color with the highest percentage is white (60.5%), black is the second most common (18.5%), while the least common is orange (2.6%) (Figure 7,

Table A8. Frequencies and percentages (%) of pictogram color.

	Frequency	Percentage (%)
Black (BK)	72	18.5
Blue (BL)	15	3.8
Brown (BR)	31	7.9
Green (GR)	12	3.1
Purple (PR)	12	2.9
Red (RD)	2	0.5
White (WH)	236	60.5
Orange (OR)	10	2.6

). Regarding the pictogram shape, the “item” category is the most populous (61.8%) (Figure 8), the second most popular category is “buildings” (10.8%) and the third most popular category is “transportation” (10.2%) (Figure 8,

Table A9. Frequencies and percentages (%) of pictogram shape.

	Frequency	Percentage (%)
Abstract design (ABST)	34	8.7
Activities (ACTIV)	11	2.8
Building (BUILD)	42	10.8
Edible items (EDBL)	7	1.8
Item (ITEM)	241	61.8
Alphabet letters (LETR)	6	1.5
Nature (NTR)	9	2.3
Transportation (TRSP)	40	10.2

). According to the semantic taxonomy of pictograms [23], most symbols exhibit a visual similarity (i.e., items, edible items, activities, nature and transportation) followed by semantic association (i.e., buildings) and an arbitrary convention (i.e., abstract design and letter).

In summary, some general trends in point symbol design are observed. Blue is the most common basic hue (26.4%) and frame background color (30.6%). At the same time, most symbols contain a frame outline (63%) which is white (24.1%). Furthermore, most symbols have a circular frame shape (65%), and most symbols are not frame pinned (62.1%). White is mainly used as a pictogram color (60.5%), and the pictogram shape (61.8%) is mostly an item.

4.2. Data Analysis per Map

The variables were analyzed for each map in the second stage. Some maps use only one basic hue for all symbols (i.e., “Visit Estonia”, “Visit France” and “Visit Norway”), while others use more than one. Most maps use more than four different hues in the symbols. Blue is the most popular basic hue only for five out of thirteen maps (Figure 9,

Table A10. Percentages (%) of basic hue for each map.

Map	Black	Blue	Brown	Green	Gray	Orange	Pink	Purple	Red	Yellow
GM	0	53	0	9	6	15	12	3	3	0
OSM	0	22	25	6	0	16	0	28	3	0
BM	0	21	0	23	3	26	0	26	0	0
MM	0	12	48	0	15	0	0	21	3	0
STM	6	17	0	17	11	28	9	0	11	25
OMT	0	11	68	0	0	18	0	3	0	0
RT	0	50	0	8	0	8	0	15	19	0
OS	0	31	24	17	0	10	7	3	7	0
HWG	0	14	8	3	0	8	37	29	3	0
SCH	0	45	0	0	0	10	20	0	0	25
VN	100	0	0	0	0	0	0	0	0	0
VE	0	0	0	100	0	0	0	0	0	0
FFR	0	100	0	0	0	0	0	0	0	0

). Seven out of thirteen maps use a frame outline for all the symbols, and three do not at all. However, three maps use a frame outline only for some symbols.

Three out of thirteen maps use a single color for their frame outline (Figure 10,

Table A11. Percentages (%) of frame outline for each map.

Map	YES	NO
GM	100	0
OSM	3	97
BM	94	6
MM	100	0
STM	6	94
OMT	0	100
RT	100	0
OS	100	0
HWG	100	0
SCH	0	100
VN	0	100
VE	100	0
FFR	100	0

). Thus, it seems that the frame outline color depends each time on the map and not on the content of each symbol. More specifically, the white color is the only frame outline color (Figure 11,

Table A12. Percentages (%) of frame outline color for each map.

Map	Black	Blue	Brown	Green	Gray	Orange	Pink	Purple	Red	White
GM	0	53	0	9	6	15	12	3	3	0
OSM	0	0	0	0	0	0	0	0	3	0
BM	3	15	0	24	0	0	0	26	26	0
MM	0	0	0	0	0	0	0	0	0	100
STM	6	0	0	0	0	0	0	0	0	0
OMT	0	0	0	0	0	0	0	0	0	0
RT	0	0	0	0	0	0	0	0	0	100
OS	0	31	24	17	0	10	7	3	7	0
HWG	0	14	9	0	0	9	37	29	3	0
SCH	0	0	0	0	0	0	0	0	0	0
VN	0	0	0	0	0	0	0	0	0	0
VE	16	0	0	84	0	0	0	0	0	0
FFR	100	0	0	0	0	0	0	0	0	0

) for “Maps.me” and “Roadtrippers”, while the black color is used for “Visit France”.

Four maps use only the circle as a frame shape (4 out of 13 maps) (Figure 12,

Table A13. Percentages (%) frame shape for each map.

Map	Speech Bubble	Circle	Pin	Square
GM	0	0	94	6
OSM	0	3	0	0
BM	0	88	0	12
MM	0	88	0	12
STM	0	100	0	0
OMT	0	0	0	7
RT	0	96	0	0
OS	100	0	0	0
HWG	0	9	91	0
SCH	0	100	0	0
VN	0	0	0	0
VE	0	100	0	0
FFR	0	100	0	0

). Google Maps and Here We Go almost exclusively use the pin, while Orange Smile uses exclusively the speech bubble. The square shape is also used several times but with low percentages and never exclusively. Since most symbols use the circle, they are not frame pinned. The maps that most often contain frame-pinned symbols are the following: Google Maps, Here We Go and Orange Smile (Figure 13,

Table A14. Percentages (%) of frame pinned for each map.

Map	YES	NO
GM	94	6
OSM	0	0
BM	0	100
MM	0	100
STM	0	100
OMT	0	0
RT	0	100
OS	100	0
HWG	91	9
SCH	0	100
VN	0	0
VE	0	100
FFR	0	100

). Most maps (nine out of thirteen) have a frame background for all their symbols. The only map that does not use at all one is Visit Norway’. Offline Map contains only seven symbols and Open Street Map only three symbols with one, while Roadtrippers has only one symbol without (Figure 14,

Table A15. Percentages (%) of frame background for each map.

Map	YES	NO
GM	100	0
OSM	3	97
BM	100	0
MM	100	0
STM	100	0
OMT	7	93
RT	99	1
OS	100	0
HWG	100	0
SCH	100	0
VN	0	100
VE	100	0
FFR	100	0

). Blue is the most frequent frame background color (Figure 15,

Table A16. Percentages (%) of frame background color for each map.

Map	Blue	Brown	Green	Gray	Orange	Pink	Purple	Red	White	Yellow
GM	53	0	9	6	15	12	3	3	0	0
OSM	0	0	0	0	0	0	0	3	0	0
BM	20	0	23	3	26	0	26	0	0	0
MM	12	49	0	15	0	0	21	3	0	0
STM	17	0	17	11	29	9	0	11	6	0
OMT	7	0	0	0	0	0	0	0	0	0
RT	46	0	7	0	8	0	15	19	0	0
OS	28	14	7	0	10	7	0	3	31	0
HWG	14	9	0	0	9	37	29	3	0	0
SCH	45	0	0	0	10	20	0	0	0	25
VN	0	0	0	0	0	0	0	0	0	0
VE	0	0	84	0	0	0	0	0	19	0
FFR	100	0	0	0	0	0	0	0	0	0

).

Four maps (Figure 16,

Table A17. Percentages (%) of pictogram color for each map.

Map	Black	Blue	Brown	Green	Gray	Pink	Purple	Red	White	Orange
GM	0	0	0	0	0	0	0	0	100	0
OSM	0	22	25	6	0	0	28	0	3	16
BM	0	0	0	0	0	0	0	0	100	0
MM	0	0	0	0	0	0	0	0	100	0
STM	6	0	0	0	0	0	0	0	94	0
OMT	0	11	68	0	0	0	4	0	0	18
RT	0	4	0	0	0	0	0	0	96	0
OS	0	14	14	17	0	0	7	7	41	0
HWG	0	0	0	0	0	0	0	0	100	0
SCH	85	0	0	0	0	0	0	0	15	0
VN	100	0	0	0	0	0	0	0	0	0
VE	0	0	0	16	0	0	0	0	84	0
FFR	100	0	0	0	0	0	0	0	0	0

) use only white as pictogram color i.e., “Google Maps”, “Bing Maps”, “Maps.me”, and “Here We Go”, while “Visit Norway” and “Visit France” use only the black color. The rest of them use more than two colors. Item is the most popular pictogram shape for all maps (Figure 17,

Table A18. Percentages (%) of pictogram shape for each map.

Map	ABST	ACTIV	BUILD	EDBL	ITEM	LETR	NTR	TRSP
GM	6	0	12	0	68	3	3	9
OSM	9	6	9	6	63	0	0	6
BM	0	3	9	0	71	3	3	12
MM	6	3	15	6	55	0	3	12
STM	6	0	14	3	63	0	3	11
OMT	7	4	4	4	68	0	4	11
RT	35	0	0	0	58	0	4	4
OS	3	3	17	0	48	10	3	14
HWG	3	0	9	3	71	0	3	11
SCH	0	0	10	0	70	5	0	15
VN	6	10	19	0	48	0	0	16
VE	3	6	16	0	61	0	3	10
FFR	41	0	0	0	59	0	0	0

). It is worth noting that this percentage is almost 50% or more for all maps.

Size depends on the visualization system. Thus, symbols on a desktop screen are larger than those on a mobile screen, so there is a broader range of values. It seems that the larger symbols are in “France.fr”, “Google Maps” and “Orange Smile“ for desktop screens. The largest size can be found in the “France.fr” map for mobile screens. In contrast, symbols with the smallest size can be found in “Roadtrippers” and “Open Street Maps” maps for desktop screens, while for mobile screens, they can be found in “Open Street Map” and “Orange Smile” maps (Figure 18 and Figure 19,

Table A19. Percentages (%) of mobile size (mm) for each map.

Map	2	3	4	5	6
GM	0	0	0	100	0
OSM	16	84	0	0	0
BM	3	0	97	0	0
MM	0	100	0	0	0
STM	0	100	0	0	0
OMT	100	0	0	0	0
RT	0	100	0	0	0
OS	100	0	0	0	0
HWG	0	10	90	0	0
SCH	0	0	100	0	0
VN	0	48	48	4	0
VE	0	3	97	0	0
FFR	0	0	0	0	100

and

Table A20. Percentages (%) of desktop size (mm) for each map.

Map	3	4	5	5.5	6	7	8	9
GM	0	0	0	0	0	0	0	100
OSM	16	84	0	0	0	0	0	0
BM	3	0	6	91	0	0	0	0
MM	0	12	0	0	66	0	0	0
STM	0	0	100	0	0	0	0	0
OMT	0	0	0	0	0	0	0	0
RT	100	0	0	0	0	0	0	0
OS	0	0	0	0	0	0	0	100
HWG	0	0	0	0	0	0	0	0
SCH	0	0	100	0	0	0	0	0
VN	0	19	29	0	42	10	0	0
VE	0	0	0	0	100	0	0	0
FFR	0	0	0	0	0	0	0	100

). There seems to be no correlation or general rule that applies to the size of the symbols depending on the visualization system. Most maps (8 out of 13) have a specific symbol size. On the contrary, different symbol sizes in the same map, i.e., OSM, Bing Maps, Visit Norway, Here We Go, Visit Norway and Visit Estonia, create the impression that some symbols are more important than others. This is not an effective cartographic practice, and it should be avoided.

In summary, some trends in point symbol design at the map level are observed. Concerning the variable basic hue, maps either use one color (e.g., Visit Estonia), or use many different colors (e.g., Bing Maps). Blue is the color that is used the most. At the same time, most maps (seven out of thirteen maps) use a frame outline in all symbols. Regarding the frame outline color, three out of ten maps (containing Frame outline) use only one color in all their symbols. Five out of thirteen use only one shape for the variable frame shape (circle (four) or speech bubble (one)). Equally, eight out of thirteen maps either use (one) or do not use (seven) frame-pinned symbols.

Furthermore, nine of thirteen maps contain a frame background, indicating the need to differentiate from the map background. Regarding the frame background color, only one map applies a single color for all its symbols. Almost half of the maps (six out of thirteen maps) use a single pictogram color (black (two) or white (five)).

4.3. Data Analysis of the Pictogram Shape

A third brief analysis was conducted, examining the pictogram shape at the map, thematic category and POI levels. At first, it was checked whether unique pictograms were used to represent each POI. In

Table 6. Percentages (%) of unique pictograms for each map.

Map	Percentage of Unique Pictograms (%)
Open Street Map	100
Sygic Travel Maps	89
Offline Map and Travel Guide	82
Orange Smile	76
Maps.me	73
Search.ch	59
Here We Go	57
Google Maps	56
Bing Maps	50
Visit Norway	29
Visit Estonia	23
Roadtrippers	11
Visit France	0

, one can see the percentage of unique pictograms on each map. Only Open Street Map uses unique pictograms, while Sygic Travel Maps, Offline Map and Travel Guide and Orange Smile exhibit a high percentage; for other maps, half of the pictograms are unique, such as Bing Maps, and also maps with smaller percentages exist. In maps that use the same pictogram for more than one symbol, these POIs tend to have similar semantic characteristics. For example, Google Maps uses the same pictogram, i.e., a bed for the symbols of guest accommodation, hostel, hotel and motel. The repeated use of the same pictogram reveals the lack of a clear POI classification. This lack leads to a deficient map as users cannot search and correctly understand the identity of points that are of interest to them. For example, if the same pictogram is used for all lodging symbols, and if the user is only interested in hotels, he cannot locate this type of accommodation when there are also hostels, motels, etc. In conclusion, it seems that using unique pictograms or not depends on the map and that most maps partially succeed in ensuring distinctiveness in point symbols.

The same analysis is applied at the thematic category level, and the calculated percentages provide more explicit results for the uniqueness of the pictograms (

Table 7. Percentages (%) of unique pictograms for each category.

General Category	Percentage of Unique Pictograms (%)
Healthcare	100
Transportation	96
Leisure–Recreation	80
Culture	73
History–Religion	70
Gastronomy–Entertainment	57
Accommodation	27
Shops	21

) and distinctiveness. The healthcare category contains only one symbol (hospital) and therefore does not allow us to draw conclusions. The low percentage of the accommodation and shops categories, which, despite the popularity on a tourist map, do not provide distinction to the different POIs and use the same pictogram, is also noteworthy. On the contrary, the transportation, culture and leisure–recreation categories are characterized by a very high percentage of unique pictograms, which is justified by the high difference in data semantics. For example, it would be crucial if the user saw the same pictogram for the airport and the bus stop on the map. In summary, there seems to be a tendency to use unique pictograms for symbols that belong in the same thematic category.

It was also examined whether a specific POI tends to be represented by a particular pictogram across maps. For example, whether for the hospital symbol the cross or the letter “H” is applied more often. In

Table 8. Percentages (%) of the most popular pictogram for each POI.

Category	Tourist Point of Interest	Most Popular Pictogram	Percentage (%)	Example of Pictogram
Leisure–Recreation	Sport activities	Man doing activity	64
	Sights	Camera	45
	Zoo	Animal	45
	Water park	Man doing activity	36
Accommodation	Motel	Bed	100
	Hotel	Bed	96
	Camping	Tent	82
	Guest accommodation	Bed	82
	Hostel	Bed	79
Culture	Library	Book	70
	Theater	Theater mask	64
	Museum	Building	69
	Cinema	Film	33
Gastronomy–Entertainment	Restaurant	Cutlery	100
	Pub	Glass	80
	Cafe	Cup	71
	Nightclub/bar	Glass	57
	Fast food	Cutlery	54
Shops	Shopping mall	Shopping bag	77
	Department store	Shopping bag	75
	Supermarket	Supermarket trolley	69
	Gift shop	Shopping bag	50
	Shoe store	Shoe	42
	Clothes store	Clothes/Shopping bag	38
	Grocery store	Supermarket trolley	38
History–Religion	Church	Church	70
	Castle	Castle	60
	Monument	Monument	40
Transportation	Bus stop	Bus	100
	Airport	Airplane	91
	Train station	Train	91
	Gas station	Gasoline pump	83
	Ferry service	Boat	73
Healthcare	Hospital	Cross	58

, for each POI, the most popular pictogram shape, the corresponding percentage and an example are presented.

POIs represented with only one pictogram are the motel (bed), the restaurant (cutlery) and the bus Stop (bus). Additionally, values greater than 80% are observed for camping (tent), guest accommodation, hostel and motel (bed), pub (glass), airport (airplane), gas station train station and ferry. Symbols that use pictograms with large percentages across maps are characterized by familiarity and acceptability, as these pictograms are generally well known and accepted by users. In most cases, these specific pictograms are also used in tourist signs [53] or road signs [54] and are thus self-explanatory to users. On the other hand, the symbol with the lowest percentage is water park, which can be considered a relatively rare POI in the maps. The more significant percentages are observed for the more popular POIs on a tourist map (e.g., accommodation category). It seems that an atypical standardization exists for popular tourist POIs. However, a tendency towards a particular pictogram does not necessarily mean that it is the most effective in terms of semantic closeness. A typical example is the gift shop symbol, which is mainly represented as a “shopping bag”, although a “gift wrapped” symbol, which is also used in some maps (e.g., OSM, Maps.me), would be more appropriate. Regarding pictogram semantic taxonomy [23], most symbols (30) exhibit a visual similarity (e.g., bed, cutlery), four a semantic association (i.e., building, church, monument, castle) and only one is an arbitrary convention (i.e., cross). Qualities such as concreteness and semantic closeness are observed.

In conclusion, using a unique pictogram per symbol depends on the map and the popularity of the category it belongs to. The frequent portrayal of a POI also seems to influence the pictogram shape. In the mind of the cartographer and the map user, a hotel is much more established as a bed than a grocery shop as a shopping cart.

5. Discussion and Conclusions

This research studied the design of tourist point symbols on web maps and apps. The aim was to investigate whether common cartographic practices are applied and whether the symbol is influenced by the POI it represents and/or by global design practices adopted by each provider. Symbols are described with a number of variables [45] (Table 3).

The first stage concerns the global analysis of symbols regardless of the map source. The sizes of symbols assure visibility and may support rapidity and ease in symbol spotting in websites and apps. Both qualities are considered important according to the bibliography [14]. Basic hue and frame background color are mainly blue. Although the blue color in traditional cartography mainly represents water bodies [55], it is also considered outside of cartography to be a strong but calm color that expands the space around it [56]. Color preference studies have also proven that the blue color is appealing [37]. In contrast, warm colors such as red, pink and yellow that exhibit small percentages are more efficient at communicating danger [57]. This is consistent with people’s preference studies that implied that yellow is the least preferred [58]. However, yellow is proposed by Medynska-Gulij as a frame background color [9]. Most symbols have a frame outline (62.8%) which is white (24.1%). This tactic provides a clear boundary and distinguishes the symbol from the map background. The white color stands out clearly and is considered an effective boundary color. Most symbols have a circular frame shape (65.1%) which is considered compact and visually stable. Circles use map space efficiently, are easy to construct and are preferred by users over other shapes. Moreover, they are compact and stable, as the eye does not wander too much over them [59]. Most symbols are not pinned (62.1%). The famous pin symbol introduced by Google Maps is only adopted by Here We Go. The use of a frame outline is also advised for the design of pictographic map symbols by a number of cartographers [9,11,14,17], but a square frame is preferred. Most symbols have a frame background (77.2%), which is also advised to help symbols pop up [11], and white is the dominant value for pictogram color (60.5%), which provides a good figure–ground relationship [17] with most background colors. Regarding the pictogram shape, the category “item” is the most frequent (61.8%). Items have high concreteness [17,22] and are therefore understandable, familiar and easy to locate on the map. Most visual variables are applied to symbol design in accordance with cartographic practices. This is consistent with a review of online map services from a cartographic perspective, which found that online map services generally conform to cartographic principles and traditional practices [60].

Some tendencies in symbol design have been noted, but it is worth investigating whether they are followed at the map level as well. Therefore, variable values are also examined for each map. The majority of the maps adopt the general tendencies in point symbol design. Blue is the most frequent basic hue in five of thirteen maps. Seven of thirteen maps use a frame outline in their symbols exclusively. Three of seven maps that exclusively use a frame outline adopt a white or black one. Four maps use mainly the circle as a frame shape. Nine out of thirteen maps have a frame background. Regarding the pictogram color, six maps use only one color for all the symbols, i.e., black for “Visit France” and “Visit Norway” and white for “Google Maps”, ”Bing Maps”, “Maps.me”, and “Here We Go”. Finally, “item” is the most popular pictogram shape. Four maps use two colors in pictograms, mostly white or black and one more.

The use of the frame outline (Table A11), the frame background (Table A15) and a specific frame shape (Table A13) is characterized by consistency and homogeneity for all maps exhibiting values close to 100%. Pictogram color is also consistent in 6 out of 13 maps. Multicolor pictograms are used with no frame background in OSM and Offline Map and in Orange Smile with a frame background. However, color is well applied only in OSM as colors group pictograms in clusters that almost follow POI classification in categories (Figure A1). The rest of the variables are applied differently for each symbol and do not support data representation. The lack of homogeneity is due to the fact that symbols are not designed as an entire set [17]. Ten maps (10/13), e.g., Bing Maps, Google Maps, Orange Smile, Here We Go and Visit Estonia, use more than one color for the frame outline. The variable symbol size in some maps questions the equal importance of all POIs (Table A19 and Table A20). Most maps (12/13) use a variety of frame background colors. In some maps, one color is used for one category (i.e., Google Maps: orange in gastronomy–entertainment, Bing Maps: red in gastronomy–entertainment, Maps.me: brown in gastronomy–entertainment and purple in shops, Sygic Travel Maps: green in accommodation and orange in shops), but this is not systematic. In most cases, colors are used randomly, are unrelated to the thematic categories and only confuse the reader. The only consistent choice is the use of red for the hospital symbol in all maps except Visit Estonia, which has a two-color symbol set (green and white). A specific frame background color or a frame outline color could be used in portraying POI categories.

A general study of symbols, even at the map level, is not sufficient since the semantics of symbols often influence their design. For example, accommodation is usually depicted with a bed, while a bag may also represent a shopping mall or grocery store. The third analysis addresses such issues, focusing on the pictogram shape. Only Open Street Map uses unique pictograms for all symbols and provides distinctiveness in the symbol set [22]. In contrast, France.fr uses all the pictograms more than once. If a tourist map aims to be detailed and clear, using unique pictograms in the symbols is more effective. Especially if the symbols belong to the same thematic category, e.g., for accommodation, the user should be able to distinguish the hotel from the hostel. At the thematic categories (e.g., leisure–recreation, accommodation, culture) level, the uniqueness of the symbols may diminish, as the POIs become similar. Characteristic examples are the accommodation category with a low share of 27% (hotel, hostel, motel etc.) and the leisure category with a high share of 80% (water park, zoo symbols, etc.). Finally, specific pictograms dominate in most popular symbols that exhibit a visual similarity such as bed for motel, cutlery for restaurant and bus for bus stop. In contrast, the less common POIs in the maps do not share a common pictogram in their symbols.

Two maps, both from National Tourism Organizations, i.e., “France.fr” and “Visit Norway”, have a specific symbol design strategy (e.g., standard pictogram color, frame shape, frame outline, frame background color and size) that makes their symbols homogeneous and recognizable. However, “Visit Norway” lags in figure–ground contrast due to the frame background absence and “France.fr” in pictogram uniqueness (0%) as four pictograms are used multiple times. As a result, it becomes obvious that one cannot judge symbol design without taking into account the pictograms’ uniqueness and representativity. On the contrary, OSM scores best in pictogram uniqueness (100%) and use of color in grouping symbols to thematic categories. It can thus be considered as the best in information transmission. As a side note, OSM is the only map that has a legend and a detailed documentation of the map features represented by symbols [61]. It can be concluded that a good symbol design requires the existence of a cartographic data ontology.

To sum up, it seems that basic hue, frame outline, frame outline color, frame shape, frame pinned and pictogram color are considered as global decisions in symbol design, while, on the contrary, the pictogram shape depends more on the symbol semantics as expected. At the same time, it should be noted that in the present study, the frame background color generally seems to vary to a greater extent than the frame shape or the frame outline’s presence, proving the importance of color in the visualization of point symbols in tourist maps. However, except OSM, the color in frame background, frame outline or in the pictogram is not used to represent qualitative aspects such as thematic categories or POI identity. In the studied maps, symbol design better addresses general map design issues such as symbol figure–ground, legibility and visual contrast that can be decided for all point symbols at a global level. In contrast, decisions at symbol level such as the shape of the pictogram in terms of uniqueness and semantics and the frame background color are less successful. POI classification in categories is highly neglected information in symbol representation.

Appropriate use, recognition and interpretation of symbols will strengthen online tourist maps in an era when travel to unfamiliar places is easier and more frequent than ever before. After all, maps are essential, and planning a trip without a tourist map is like building a house without drawings [62]. Symbol standardization in tourist maps as in topographic maps can be considered as solution [9,15]. First Ratajski (1971) [19] has proposed a standardized set of symbols for economic maps and school atlases. Standardization is also related to familiarity, which is also a decisive factor [12]. Standardization can be implemented through an iterative collaboration process [63]. Nowadays, symbol standardization is a topical issue in crisis maps [27,64]. However, in the case of tourist maps, it is the original graphic design that makes them attractive for the users [9]. A possible solution is the creation of templates or tools that can guide the cartographer in designing symbols for a tourist map that fully express semantics with visual variables and are in accordance with cartographic best practices and guidelines. This guidance would support symbol design without excluding creativity and innovation. In the future, research on pictorial point symbol design for tourist maps should focus on the user aspect. Information on how map users interpret and best understand symbols [65,66] is essential to improving symbol design, especially user’s preferences of pictograms.