You are on page 1of 21

International Journal of Cartography

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tica20

A participatory trail web map based on open


source technologies

Joshua Gore, Stefan Peters & Delene Weber

To cite this article: Joshua Gore, Stefan Peters & Delene Weber (2022) A participatory trail web
map based on open source technologies, International Journal of Cartography, 8:2, 223-242,
DOI: 10.1080/23729333.2020.1840320

To link to this article: https://doi.org/10.1080/23729333.2020.1840320

Published online: 08 Feb 2021.

Submit your article to this journal

Article views: 144

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=tica20
INTERNATIONAL JOURNAL OF CARTOGRAPHY
2022, VOL. 8, NO. 2, 223–242
https://doi.org/10.1080/23729333.2020.1840320

A participatory trail web map based on open source


technologies
Joshua Gore , Stefan Peters and Delene Weber
University of South Australia, UniSA STEM, Adelaide, Australia

ABSTRACT ARTICLE HISTORY


Interactive maps can be an important marketing tool for Received 19 May 2020
disseminating information about long distance walking trails and Accepted 19 October 2020
a way to value add to the recreation experience. Providing
KEYWORDS
participatory functionality to maps by allowing users to share Web map; open source;
new information or refine existing information can greatly participatory; volunteered
enhance user interest and improve the product. This research geographic information;
uses open source front and back end technologies to develop a tourism
single page, database driven participatory web map application
for the Walk the Yorke Trail in South Australia. The development
is based on and assessed through a user-centred design
approach. The prototype participatory web map is assessed by
cartographic experts, trail managers, and trail users. These
evaluations indicate the validity of the design directions taken
but highlight the need for information quality and quantity when
encouraging participation from knowledgeable trail users, and
the need for tools to ensure the continuing quality of further
submitted information.

ABSTRACTO
Les cartes interactives peuvent être un outil marketing important
pour diffuser des informations sur les sentiers de randonnées de
longue distance et un moyen pour donner de la valeur ajoutée à
ce loisir. Proposer une fonctionnalité de participation aux cartes
en permettant aux utilisateurs de partager des nouvelles
informations ou d’améliorer les informations existantes peut
considérablement augmenter l’attrait de l’utilisateur et améliorer
le produit. Cette recherche utilise des technologies open source
pour développer une application de carte web participative
présentée sur une page unique et basée sur une base de
données pour le sentier Walk of the York en Australie du Sud. Le
développement est basé et validé par une approche de
conception centrée utilisateur. Le prototype de carte web
participative est évalué par des cartographes experts; des
gestionnaires de sentiers et des randonneurs. Ces évaluations
confirment la validité des choix en matière de conception mais
soulignent le besoin de qualité et de quantité de l’information
pour encourager la participation de randonneurs avertis et le
besoin d’outils pour assurer la qualité continue des informations
supplémentaires soumises.

CONTACT Stefan Peters Stefan.Peters@unisa.edu.au


Supplemental data for this article can be accessed https://doi.org/10.1080/23729333.2020.1840320.
© 2021 International Cartographic Association
224 J. GORE ET AL.

Introduction
A web presence is essential for advertising tourism opportunities in today’s world (Efte-
khari et al., 2011; Pack, 2019). Similar to the traditional role of word-of-mouth, Web 2.0
participatory applications, such as online blog forums and review platforms, are signifi-
cant channels for advertising tourism opportunities (Confente, 2015; Saura et al., 2020;
Yoo & Gretzel, 2011).
Tourism, or travelling for pleasure, is by definition spatial. Thus it is not surprising that
many tourism websites incorporate some form of web map. Consumer-focused map plat-
forms such as Google Maps demonstrate the popularity of Web 2.0 participation, allowing
users to add information (reviews, comments, and questions) to existing features and in
some cases contribute new spatial data. However, despite claims of potential, little
research has been conducted into how such functionality can form part of an effective
tourism-focused cartographic solution. A well-designed product will not only improve
awareness but could significantly improve the overall visitor experience.
Technologies which have emerged over the last two decades permit cartographers to
create dynamic web map applications which allow extensive interactivity. Vector tiles and
WebGL accelerated vector tile viewers have greatly enhanced performance, reducing
download requirements and enabling extensive cartographic customisation (Eugene
et al., 2017; Nordan, 2012; Yao & Li, 2018). Binary vector tile technologies also provide a
performant interface to spatial databases without the overhead of uncompressed json-
based formats (Eugene et al., 2017). When combined with single page web application
libraries, these new technologies allow performance and functionality approaching that
of traditional installed applications (single page web applications are sites which down-
load information asynchronously to dynamically change a single page rather than requir-
ing users to load new pages for further information).
This research explores the application of open source database driven single page appli-
cation technologies and vector tiles to the development of web maps designed for aiding
long distance trail-based tourism. The research uses a user-centred design approach to
develop and assess a web map for the recently created (officially opened on the 13th December
2015) 500 km Walk the Yorke Trail along the coast of South Australia’s York Peninsula (see
Figure 1). The trail provides a diversity of experiences and routes for both hiking and cycling
and is in close proximity (between one and three hours) to the state’s largest tourism market

Figure 1. Map of trail location.


INTERNATIONAL JOURNAL OF CARTOGRAPHY 225

and capital, Adelaide. Whilst many sections of the trail have limited mobile reception coverage is
available at the townships typically marking the start and end of sections. Research was there-
fore primarily focused on tasks users are likely to perform at these locations before and after trail
usage, with the use of GPX files and paper maps encouraged for navigation purposes.

Literature review
Existing research into digital trail cartography centres around the potentials of mobile
mapping (e.g. Hauthal & Burghardt, 2012; Kessler, 2011; Nivala et al., 2009; Sarjakoski
et al., 2012; Sarjakoski & Nivala, 2005). The majority of general mobile cartographic
research focuses on Location-Based Services (LBS), Volunteered Geographic Information
(VGI), and adaptive cartography (Roth et al., 2018). Trail cartographic research likewise
centres around these areas.
A contextual cartographic approach implemented using web technologies may use all
three of these areas along with mainstream web design approaches to provide a web
map which adapts to suit changing contexts. A contextual model is necessary to create
effective contextual adaptions and ensure appropriate interactive functionality. In their
review of contextual research progress, Griffin et al. (2017) suggested a model in which a
map’s context is defined as the interaction between four key areas: the user (abilities, dis-
abilities, and cultural differences), user’s activity (what is being done and when), map usage
environment (distractions and physical factors such as bright sunlight or weather) and the
map itself (in a web context, the diversity of devices that can be used to view a map).
Modern technology has fundamentally changed cartographic models Hauthal and Bur-
ghardt (2012). The role of cartographers and users has morphed, with users becoming
mapmakers by both contributing data and customising maps to suit their needs. Carto-
graphers, on the other hand, provide interactive functionality and make expert choices
about the extent to which users can change cartographic representations. Such choices
can be guided by research. Nivala et al. (2009) identified 101 trail user needs and distilled
these into the nine categories shown in Figure 2.
Another form of contextual adaption used in mainstream web development is responsive
web design. Responsive web design describes design approaches that allow websites to
adapt to the wide variety of client devices in use today, in contrast to sites with separate
mobile and desktop pages (Ricker & Roth, 2018). The application of responsive design to car-
tography is an emerging area of cartographic research, which includes adaptions to users’
individual abilities, activities and environments (Griffin et al., 2017; Ricker & Roth, 2018).
Prior to the emergence of smart phones, Sarjakoski and Nivala (2005) provided sugges-
tions for how cartographic contextual adaptions can be practically implemented on
mobile hiking navigation devices. Based on users’ assessments of existing maps, they
identified areas where contextual adaption may be necessary and classified as active or
passive. Active adaptions are changes in response to automatically sensed contexts,
such as location and time of day. Passive adaptions are changes in response to additional
information provided by users, such as their age, cultural background and preferred
activities. Sarjakoski and Nivala (2005) gathered this information through providing an
initial menu gathering such information, and then used the outcome of this menu to
set language settings and display relevant information in a cartographic style specific
to the target audience.
226 J. GORE ET AL.

Figure 2. User requirement categories identified by Nivala et al. (2009).

Dominguès and Hoarau (2011) used an adaptive approach to design paper trail maps
for several trail user groups. They used a combination of surveys and interviews to
develop four distinct map styles for casual walkers, dedicated hikers, cross-country cyclists
and road cyclists. Key cartographic questions they recommend should be addressed
when developing adaptive maps are:

(1) What is the technical ability of the group and how should this influence design
complexity?
(2) Which touristic points of interest are relevant to the user group?
(3) What information scale is most useful for the group?
(4) What activity-specific information does the group require?

This user-centred design approach is commonly used for software development.

User-centred design
The process described by Norman and Draper (1986) in their seminal work on user-
centred design has become an important part of software development, encapsulated
in ISO standard ISO 9241-210:2010 (ISO, 2009). This process is an iterative four step loop:

(1) Understand the setting in which the application will be used


(2) Determine users’ needs
INTERNATIONAL JOURNAL OF CARTOGRAPHY 227

(3) Develop prototypes meeting these needs


(4) Evaluate prototypes (improving developers understanding of context and user needs)

It was not until the 2000s with the release of web map applications such as Google
Maps that the value of this process was recognised for creating digital maps, a shift in
focus from providing functionality to experts to providing information tailored to assist
the users (Tsou, 2011).
Roth et al. (2015) provide a comprehensive overview of user-centred design
approaches and detail the specific considerations and processes necessary in a web
mapping context. They propose a three-part loop used throughout the design progress:

(1) User: what are the users’ needs and capabilities?


(2) Utility: how can the functionality of the application meet these needs?
(3) Usability: how easy to use and attractive is the application?

Applications typically need to compromise between utility and usability. For example,
an application may be very well designed and simple but provide few functions or may
provide many functions but be cluttered and hard to use. Roth et al. (2015) describe
three areas of evaluation:

. Expert-based methods: evaluating in collaboration with usability and domain experts


. Theory-based methods: evaluation by designers and developers based on existing
usability theories
. User-based methods: feedback from a representative group of users, both formative and
summative

Methodology
Following the methodology outlined by Roth et al. (2015) the project development fol-
lowed a single UCD cycle, concluding with both a user and an expert based assessment.
Unlike the case study around which Roth et al. (2015) based their suggestions, the project
focused on implementing functionality that would allow the wide range of functionality
and information requested by users rather than on collecting and presenting further infor-
mation to users. The methodology was designed to provide specific suggestions around
design and interactions which can be used to create future web maps. Table 1 outlines the
specific tools and methods used to develop and assess the prototype web map app.
Electronic and paper questionnaire surveys were the primary tools used in the first and
second ‘user’ stages due to the wide-ranging locations of both the trail and trail users, and
in order to allow assessment of the application by international cartographic experts. At
the time of development, Mapbox Vector Tiles and associated APIs provided the most
functionality and cartographic flexibility. Other technologies were chosen based on cur-
rency and ease of development.
Hotjar Analytics tracking (Hotjar Ltd, 2019) was used to gain a deeper understanding of
user behaviour. Information from Hotjar was not used as a primary assessment method
but aided understanding of difficulties noted by survey respondents through providing
logs and a heatmap of user map interactions.
228 J. GORE ET AL.

Table 1. Overview of project methodology.


Stage Process Method/Tools
1 (User) Initial gathering of trail user information and Trail Users: Electronic and paper survey
trail manager desires Trail Managers: Project meetings
2 (Utility) Design database storage and access APIs. Relational Database: PostGIS
GraphQL API: Hasura
Authentication: Node/Passport
Vector Tiles: Tegola
3 (Usability) Develop front end interface GraphQL: Apollo Client
Data Management/Interface: React
Map: Mapbox GL JS
4 (User) Assessment of interface usability and utility Expert and Stakeholder Assessment: Task-based Survey,
Hotjar Website Analytics
User Assessment: Task-based Online Survey, Hotjar
Website Analytics

User: initial research


Trail managers
The main body in charge of promotion and management of the Walk the Yorke trail is the
Yorke Peninsula Council. Several discussions were held with members of the Council
involved in tourism and trail management. These discussions, summarised in Table 2,
identified the potential value user submitted information could have to the Council, high-
lighted the need for the web map to showcase the trail opportunities in an engaging
manner, and explored some of challenges around a long-distance remote trail. Council
members centred discussion around possibilities more than limitations and concerns of
Web 2.0 technologies.

Trail users
A survey was distributed targeting current and potential users of the Walk the Yorke Trail.
These individuals were either approached when engaged in trail usage, identified by the
Council, or members of local organisations who promote outdoor exercise. The survey
explored trail usage, digital and paper map preferences, technical abilities, demographics,
and suggestions on information and functionality. These suggestions, summarised in
Table 3, drove the UCD cycle undertaken as part of this project.

Table 2. Summary of functionality suggested by trail managers.


Feature/Information Details
Trail Registration Few statistics on trail usage: incentivised trail registration system to research usage and
improve safety
Alerts and Hazards – Web map should include known, ongoing alerts and hazards
Ongoing
Alerts and Hazards – Web map should be updatable with new, time specific alerts and hazards
Temporary
Emergency and Policy Web map should be able to effectively communicate emergency contacts and trail
Information legalities
Reviews and Information Web map should encourage facilitation of trail communities
Sharing
Social media integration Web map may integrate with existing social media pages/hashtags
Water Sources Users should be able to share specific information, particularly potential water sources
and current water storage container levels
Application maintenance Council staffing is limited with no dedicated GIS staff thus map needs to be easy to
maintain
INTERNATIONAL JOURNAL OF CARTOGRAPHY 229

Table 3. Summary of information and functionality suggested in user survey.


Cyclists Walkers Hikers*
Feature/Information (10) (23) (13) Total
Potable Water 4 5 5 14
Trail Surface 5 1 4 10
Bitumen, dirt, sand, gravel, natural rock, hard/soft, road or track, road
bike traversable
Elevation/Grade Profile 3 4 1 8
Scenic and historic points 2 4 2 8
Distances 2 3 3 8
Between major points/towns/campsites
Camping 2 4 2 8
Parking and Access 1 2 3 6
GPX Download 1 2 2 5
Toilets 2 2 1 5
Facilities 2 3 5
Bike Supplies, Food, other shops
Accommodation 1 3 1 5
Paper Maps Preferred 2 2 4
Difficulty 3 1 4
Routes, Walks, Half/Full Day Sections 2 1 3
Cultural Information 2 1 3
First Nations Names/History
Usage/Users (e.g. vehicles and trail bikes) 1 1 2
Beach Safety and Activities 1 1 2
Trail Sign Locations 1 1 2
Satellite 1 1 2
Gas/Wood Fire Legality 1 1 2
Environmental Information 1 1 2
Native and Invasive Flora and Fauna
Accuracy of Data Emphasised 1 1 2
Trail Width 1 1
Fences/Gates 1 1
User Reviews 1 1
Mobile Reception 1 1
Farm Boundaries/Private Property 1 1
Emergency Contacts 1 1
*Hikers are users who complete overnight or longer walks.

To aid discussion the survey included a link to a web map demonstrating basic adaptive
cartography and participatory point submission, as suggested by Slocum et al. (2003). The
survey was distributed purposively in both hard copy and digital formats to trail user
groups, with a focus on users of the Walk the Yorke Trail. To ensure a wide range of perspec-
tives the opinions of a few users unfamiliar with the trail were also sought with users asked to
rate both their familiarity with the Yorke Peninsula and with the Walk the York Trail. The
summary in Table 3 shows most suggestions received related to information rather than func-
tionality. Survey documents can be accessed at https://doi.org/10.5281/zenodo.4046031.

Utility: information storage and access


Users requested information about potable water, trail surfaces, elevation and points of
interest. However, as the trail has just been launched and information is still being compiled
and curated, it was not possible within the constraints of this project to fulfil all demands. A
subset of information and functionality was therefore identified based on existing data with
the intent of demonstrating the technical feasibility and cartographic potential of imple-
menting a greater range of suggestions. This subset is summarised in Table 4.
230 J. GORE ET AL.

Table 4. Subset of suggested functionality implemented in prototype web map.


Feature/Information Details Functionality demonstrated
Routes Provided trail linear information was divided into . Line feature interaction
individually selectable stages with related textual . Symbolisation by line type
. Bike
information . Retrieval and display of textual
. Hike
. Shared Use information

Distances Route distances and travel times were included in . Retrieval and display of textual
descriptions information

Points of Interest Trail shelter data was available thus was used to . Point feature interaction
demonstrate interactive points of interest with . Retrieval and display of
. Trail Shelters
associated multimedia multimedia

Non-Interactive Available point data was displayed on the map, and . Storage and display of large
Information symbolised based on zoom level (1000+) numbers of points
. Scale-dependant and contextual
. Toilets
. Signs cartography
. Seats
. Markers

Alerts and Hazards – Point hazards and issues were able to be viewed and . Point-Based VGI
Temporary submitted through front end interface . User friendly point submission

Reviews and Point submissions of campsites, water sources, issues, . VGI related to existing spatial
Information Sharing and hazards were allowed. All user points, interactive features
points, and routes can have related comments and . VGI point collection
. Comments and
Reviews reviews.
. Submission of
Points

Application Standard spatial database used . Spatial information modifiable


maintenance using standard GIS software
. GUI-based management of
related information

GPX Download All stages have related GPX files . Storage and retrieval of files
related to features

As many sections of the Walk the Yorke trail are remote, with limited mobile recep-
tion the application was primarily built around tasks that the user was likely to perform
before or after trail usage, similar to Nivala et al. (2009). GPX downloads for use in exist-
ing applications and links to purchase topographic maps were included to aid users
whilst on the trail.

Database design
Based on user suggestions and available data, the normalised database structure shown
in Figure 3 was used to store the information shown in Table 4. Normalising spatial
data and related information greatly enhanced the potential utility of the database,
allowing for addition of further and refinement of existing information without the
need for modifying the database structure or duplicating spatial data. This structure
was implemented using the PostgreSQL database system with the PostGIS extension
as this combination provides excellent performance, allows numerous integrations
with other web technologies, and is accessible by commonly used Desktop GIS
packages.
INTERNATIONAL JOURNAL OF CARTOGRAPHY 231

Figure 3. Diagram of web map database structure (pivot tables have not been included).

Firstly, a many-to-many relationship between a categories table and other attributes


was used. This ensured standardisation of categorised feature properties, essential for
correct cartographic symbolisation. A self-relation within categories allowed grouping
and future refinement of existing categories. This was used to define whether a user
point was a new feature or an issue, with sub-categories for each.
Secondly, lines were grouped by routes which in turn could belong to other routes.
This approach allowed classification of routes as part of a trail network (i.e. the Walk
the Yorke Trail) and grouping of lines into stages. The approach allowed data to be
further split to reflect changes in surfaces and key navigational waypoints or updated
to reflect changes in the trail route without affecting the structure of information
served to the frontend client.
Finally, links to file downloads and multimedia were stored in separate tables with
many to many relations to routes and interactive points and a basic user role table was
added to allow identification of administrators, moderators, and trail managers.
232 J. GORE ET AL.

Database services
To allow the utility of the database structure to be accessible to a front-end interface
two key database APIs were implemented. The first, a GraphQL-based JSON API, pro-
vided data as objects and arrays, necessary for providing a variety of related infor-
mation for functionality built around the map. In contrast to more traditional API
services, GraphQL allows retrieval of multiple related objects in a single query, reducing
network usage and simplifying development. The second, a Mapbox Vector Tile API,
provided spatial data and attributes in the flat structure necessary for cartographic rep-
resentation. Whilst spatial data could be retrieved using the GraphQL API, the vector
tile API resulted in a noticeable performance increase and greatly simplified develop-
ment through eliminating the need to pre-process data into a web map compatible
format.

Usability: information access interface


Application development
The open source React Java Script library was used for application development due to its
broad ecosystem of development tools and libraries. Application content was primarily
loaded from the database, with properties such as the trail to load, site name, and API
URLs set through a.env file. This abstracted approach prevented duplication and allowing
flexibility in use. The application described in this article can be accessed at bit.ly/
WalkTheYorkeMap.

Application design
To translate application utility into a usable interface, functionality was grouped into two
areas, following the approaches of the two database APIs:

(1) Multimedia, files, user details, detailed textual information, comments, and reviews
(2) Spatial information and related details necessary for cartographic design

The application utilised a panel for display of non-spatial information, displayed under-
neath the map on mobile devices and beside the map on larger devices. This panel could
expand to fill the entire display, allowing presentation of detailed information as would be
done on a traditional website. Map interactions were the primary method of controlling
the information within this panel. This two-panel system enabled logical separation
between spatial information, loaded on map load and zoom, and more detailed information
loaded from the database on feature selection. Information displayed on the map itself was
kept to the minimum needed for identification of features to ensure fast loading. Loading skel-
etons were used on the side panel to make the slower loading times more acceptable to users.
All information shown in the side panel was directly related to the map selection.
Auxiliary functions were provided through buttons located in the top corner of the
map. These linked to temporary modal menus which provided authentication, display
of help information, and access to a participatory point submission mode.
The main application could be viewed in the four states shown in Figure 4. On initial
load, the application showed a view of the entire trail (3a), with associated description and
downloads. Tapping or clicking on a trail section in the map selected the stage mode (3b).
INTERNATIONAL JOURNAL OF CARTOGRAPHY 233

Figure 4. Web map design.

When this mode was selected, the map zoomed to the selected stage route. The selected
stage route was highlighted through an increase in line width, and the rest of the trail de-
highlighted as dashed lines. Users could return from this mode to trail mode through
either a ‘Home’ breadcrumb link shown on the panel or through tapping an area of
the map with no interactive objects. Tapping or clicking on an interactive point
entered point mode (3c). The point icon was replaced with a larger red balloon icon to
show which point was selected, and again information was loaded about the selected
point. Users could return from a point selection by tapping on an area of the map with
no interactive objects or selecting another object. Finally, selecting an option from the
submit point button in the corner of the map entered new point mode (3d).

Participatory functions
The new point mode was one of two forms of participation included in the web map. The
function allowed users to submit both unrecorded physical features such as camping
spots and water sources and emerging trail issues such as hazards and damage. The
other form of participatory functionality included was refinement of existing information
through comments and reviews. Comments and reviews could be submitted for all inter-
active routes, points, and user submitted points. Reviews allowed users to share their trail
experiences through submitting a text and rating out of five, with an average rating
234 J. GORE ET AL.

shown under the title in the information sidebar. The comment interface allowed users to
submit and reply to comments. Top level comments could be tagged as issues through
starting with the text ‘#issue’, stored as a related category in the database. This function
was intended as a proof of concept for improving comment-style participatory function-
ality, with the potential for expanding to other tags such as #question. These tags could
be used to create a forum-style overview of comments encouraging participation from
knowledgeable trail users and notifying trail managers of potential issues. The
comment interface allows users to contribute to user-submitted points without duplicat-
ing information.

Cartography
Unlike traditional raster tile basemaps, Mapbox vector tile basemaps allow integration of
user data within the basemap layer stack. A modified version of the Mapbox Outdoors
style was used as the basemap for the trail web map. To enable the use of multiple base-
maps without the need to define insertion points for each, a transparent placeholder layer
was inserted above basemap roads but below road labels and points of interest. This
approach resulted in an integrated appearance and ensured basemap information was
readable.
The Mapbox style specification also simplifies styling based on map scale. Filters are
a common form of passive adaption provided on trail maps, allowing users to limit
information shown to aid readability. Scale-dependant filtering was used as an alterna-
tive active adaption, with shelters appearing at a trail scale, user-submitted points
appearing at a stage scale, then progressively toilets, signs, benches, and trail
markers. Labels were likewise scale-dependant, appearing at higher zoom levels than
their respective point symbols to ensure readability. This approach allowed display of
over 1000 initial points and potentially many more user-submitted points without
the need for any user input.
The map symbology used is shown in Figure 5. As shown, a split-complementary
colour scheme based on the trail logo was used for lines, with the complementary blue
colour used for cycling sections as these were often more accessible detours following
roads adjacent to the trail. Planned route usages were also symbolised by icons placed
along the trail lines, an approach intended to both make colour symbology clear and
to improve accessibility. Non-interactive point symbols were designed to harmonise
rather than contrast with the Mapbox basemap. Interactive points were designed to con-
trast through using vivid colours and were represented by the balloon shape commonly
used to indicate interaction. As user-submitted point types could be changed through the
database, they were only symbolised as either points or issues. The specific type of point

Figure 5. Web map symbology.


INTERNATIONAL JOURNAL OF CARTOGRAPHY 235

as well as the submitted point name were however shown as the point label. To identify
the source of these points and allow for submission of points from official sources the text
‘user submitted’ was also included in labels.

User: application assessment


Two internet-based questionnaire surveys were distributed in order to assess users’ atti-
tudes towards the applications utility and whether this utility was implemented in a
usable way. The first was distributed to cartographic experts, tourism experts, and individ-
uals involved in the promotion and management of the Walk the Yorke Trail (referred to
as experts or expert evaluators). The second was distributed to current and potential users
of the trail, primarily those who were involved in the initial user survey used as the basis
for development (referred to as user evaluators).
These surveys were composed of a series of quantitative and qualitative questions primar-
ily based around a series of tasks exploring the design and functionality of the web map app.
As shown in Table 5, the first half of these tasks explored the map’s cartographic design and
associated feature interactions, and the second half the map’s participatory functionality. The
surveys sought both formative information on flaws in the application which reduced usabil-
ity and summative information on the success of overall design and functionality.

Results
Table 6 shows the tasks evaluators completed and summarises the overall usability
ratings for each task. Expert evaluators completed six tasks and user evaluators four

Table 5. Web map evaluation tasks.


Tasks – cartography Purpose Expert User
View a description of a trail stage . Assess Line Interactions X X
. Assess Textual Information Layout
. Demonstrate Adaption from trail to stage context

Find a toilet and an information sign . Assess scale-dependent cartography X


. Assess cartographic Symbology

View information associated with a trail shelter . Assess point interactions X X


. Assess balloon symbol used for interactive points

Tasks – participation Responses Expert User


Review a stage . Assess review interface X X
. Gather views on value of participatory functions

Submit a comment with a tag . Assess comment interface X

Add a point . Assess spatial point submission functionality X X


. Gather views on spatial participation

*1–5 scale where 1 = very hard and 5 = very easy.


236 J. GORE ET AL.

Table 6. Ease of completing web map tasks as perceived by experts and users.
Expert rating User rating
Cartography Responses Average* Responses Average*
View a description of a trail stage 22 4.3 16 4.6
Find a toilet and an information sign 22 3.7
View information associated with a trail shelter 21 3.9 15 3.9
Participation Responses Average Responses Average
Review a stage 16 4.6 11 4.5
Submit a comment with a tag 15 4.3
Add a point 14 4.8 10 4.3
*1–5 scale where 1 = very hard and 5 = very easy.

tasks. Questions relating to these tasks were reduced and simplified for evaluators in the
user group. Ratings provided support the design decisions described, however several
usability bugs were identified as were several potential utility enhancements. The main
issues encountered were around feature interactions, particularly when evaluators
attempted to interact with a non-interactive point near an interactive trail line.
Evaluators were asked whether they would prefer scale-dependent filtering, filtering
through an interface, or a combination of the two. Table 7 summarises the results of
this question. At the start of the survey, evaluators were asked which devices they
were using to complete the assessment. Whilst the majority of experts who indicated
they were using desktop or laptop computers would prefer filter through interface con-
trols (56%), most experts (80%) completing their assessment on a smartphone preferred
the scale-based approach.
Results from the six user tasks are described below.

View a description of a trail stage


Two main issues arose with the method used to allow interaction with trail stages in order
to view related information. The first, relating to the overall map interaction approach,
was that some evaluators would zoom into the trail and then click on a stage, causing
the map to zoom back out to a more general context. This issue could potentially be
resolved by extending the map’s scale-dependent symbology approach to include
feature interactions, either disabling line interactions past a certain scale or just prevent-
ing zooming out. The second was that some evaluators would mistake stage zooming for
interactions with non-interactive points, particularly as clicking on a non-interactive area
of the map cleared the selected stage and zoomed to the trail’s extent. Preventing inter-
action with lines near non-interactive points would help resolve this issue. A related issue
was that when a trail section selection was cleared by clicking in an area of the map with
no interactive objects, the map would zoom out to a view of the entire trail. Whilst clear-
ing a selection by clicking elsewhere would be an expected interaction, zooming to the

Table 7. Scale-based vs interface-based filtering.


Interface filtering Scale-based filtering Both
All 10 45% 7 32% 5 23%
PC 10 56% 3 17% 5 28%
Mobile 1 20% 4 80% 0 0%
INTERNATIONAL JOURNAL OF CARTOGRAPHY 237

entire trail would only be expected when specifically selecting this context. This confusing
interaction should be removed as the ‘Home’ breadcrumb link provides the same func-
tionality through a more predictable interface.

Find a toilet and an information sign


A key variable predicting negative evaluations was expectations. When users expected to
be able to interact with a non-interactive point, or experienced unexpected interactions
due to nearby trail lines described in the previous section, their rating was lower.
The evaluation of these non-interactive point features highlighted key considerations
when using a scale-dependent symbolisation approach. Points can be considered either
as related to the current context of a web map – symbolised to provide contextual infor-
mation around what the user is viewing, or as the direct focus of the user’s view – as
required by this task. Determining which approach to use for each point ultimately
depends on the amount of information presented and the tasks the map is intended to
allow. The map could potentially be expanded to allow map users to select filtered infor-
mation to be shown at all zoom levels whilst defaulting to a scale-dependent contextual
representation usable by those on mobile devices. Typical web basemaps already provide
scale-dependent symbolisation. A cartographic design intended to integrate with
basemap symbology must draw a distinction between information specifically related
to the trail but intended to be part of the basemap context, and both interactive and
non-interactive features critical to high-level planning.

View information associated with a trail shelter


This task explored the symbolisation of interactivity through balloon symbols. As with
the previous task, several assessors indicated they expected interactivity based on the
feature type rather than the feature symbol. One additional feature suggested for inter-
active features was on-map popups triggered either on point selection or on mouse-
over. Whilst popups are commonly used on web maps, they were not included on
the Walk the Yorke web map as the side panel was considered more mobile friendly.
Mouseover interactions were likewise largely avoided to ensure implemented map
interactions would work on all devices. However as suggested by one evaluator
popups could help identify points in place of lengthy labels with the side bar providing
more detailed information.

Review a stage
Most evaluators indicated that the participatory functionality would be useful but that
functionality for moderating submitted information and removing old information relat-
ing to resolved problems would be essential. Several solutions ranging in technical
difficulty were proposed. A simple improvement to the review interface would be to
include the date reviews were submitted, however a more complete participatory sol-
ution would allow users to upvote useful information, report irrelevant comments to
moderators, and allow trail managers to close or remove issues.
238 J. GORE ET AL.

Another participatory improvement suggested was allowing map users to share


photos either directly through the application or by integration with social media. Allow-
ing photo upload to the application would be straightforward but the increased storage
cost may outweigh the benefits of allowing this functionality. Further research is needed
into users’ attitudes towards sharing photos on general-purpose social media sites with
tags intended to associate their posts with a custom application. Whilst trail users were
already sharing photos tagged to the Walk the Yorke trail on Facebook and Instagram,
associating these photos with specific trail objects would require users to use a much
more granular approach and to include information not necessarily of interest to their
social media audiences.

Submit a comment with an issue tag


In addition to the concerns already discussed around management of submitted com-
ments and reviews, the main concern raised with this function was integration with exist-
ing management policies and systems. The Council had an existing customer relationship
management system. Council policies required requests for repairs to specific trail assets
to be submitted through this system’s existing online form. Some evaluators also indi-
cated documentation was needed for map users to understand what tags meant and
that further tag categories would be valuable, however issues with integration and man-
agement by Council would need to be resolved before this could be provided. A menu
rather than text-based tagging system may also have enhanced usability.

Add a point
The ease of completing this task was rated the highest of all tasks from expert reviewers,
however some user evaluators suggested that the side form should have been made
more explicit. Other improvements suggested included allowing uncategorised points
through an ‘other’ category, providing further details on what each point was intended
to represent, and including a date on submitted points to help determine relevance.
There are obviously trade-offs in creating an ‘other’ category and it suggested if this is
done, it is done for a specific exploratory period and then the data should be analysed
to see if the points could have been classified into existing categories or if indeed impor-
tant categories were missing. Again, evaluators highlighted the need for data moderation
and removal of old data.

Overall comments and ratings


Two key additional feedback items came out across multiple tasks and in general
comment boxes. The first, particularly noted by user evaluators from the Yorke Peninsula,
was that their ratings of the map and inclination to contribute to the map was dependent
on the quality and quantity of information presented. The intent of this project was to
explore functionality which could present a wide range of information however as has
already been noted the Walk the Yorke trail has only been developed recently thus
much of the information requested by users is not yet available. This feedback also high-
lights a key trade-off in the design of participatory trail web maps. Increasing the quantity
INTERNATIONAL JOURNAL OF CARTOGRAPHY 239

Table 8. Overall web map ratings.


Experts Users
I think I would like to use the web map when planning my trail use 8.4 9
I though the web map was easy to use 8.4 9
I thought the various functions in the web map were well integrated 8.4 8.3
I would imagine that most people would learn to use the web map very quickly 8.4 8.9
*1–10 scale where 1 = strongly disagree and 10 = strongly agree.

and complexity of information in a map will help increase contribution of further data
from knowledgeable users, however with increasing complexity comes a need for
increased utility, potentially reducing usability and thus discouraging users who just
want a simple map to overview the trail. The effectiveness of any change or design
choice is made much easier by keeping the key target audience in mind.
The second was feedback around usage of the web map on the trail. Several evaluators
suggested routing or other navigation functionality. A couple of evaluators also
suggested the trail map would not compete with currently available topographic maps.
Due to poor mobile reception along sections of the trail, the map was not actually
intended to be used as a replacement for a topographic map, though GPX files were pro-
vided to be used with popular trail map applications or GPS devices. The comments do
however reflect the increasing dependence and demand for phone-based apps. Many
people will not bring a paper map if a map is freely available online. For a trail web
map to be successful, both the intended context and the contextual limits of the map
must be made clear to users with links to other solutions for on-trail contexts, particularly
in areas with extreme conditions and limited trail infrastructure.
Table 8 summarises the overall ratings given by both expert and user evaluators. These
ratings indicate the overall map design approach was effective, particularly if the dis-
cussed improvements were implemented through further UCD cycle iterations.

Discussion
Whilst summative ratings showed the design and functionality directions taken in creating
the prototype Walk the Yorke participatory web map were effective, feedback also indi-
cated a number of challenges that will be encountered in developing participatory maps
for long distance trails, and in developing single-page applications designed to adapt to
both mobile and desktop devices. Consideration of contextual design and adaptions for
web maps must include analysis of interactions at varying map scales and how these inter-
actions may change across touch and non-touch devices. A web map application’s non-
spatial interface elements must also form a part of an effective contextual design.
Textual information and menu-based navigation are important both for users wanting to
choose a section of a trail to complete that suits their tastes and for aiding participation
through showing a forum style overview of recently shared questions and issues.
The utility suggestions provided by Nivala et al. (2009) are now not only technically
possible but frequently provided by mainstream smartphone applications. Features
such as sharing experiences along with photos, sharing one’s current location with
friends, recording and sharing hikes and rides, and offline navigation are all popular
tasks performed by mobile apps. Care must therefore be taken to ensure features
240 J. GORE ET AL.

made available in new trail web maps, particularly those designed for specific trails, do not
duplicate features popular in mainstream applications unless significant value is added.
Users may be more likely to share photos and experiences with friends using mainstream
social media applications than through a specialist application catering only to a specific
group. Integration with mainstream applications, particularly social media, is therefore
important, potentially greatly enhancing the value of user participation both for users
and for organisations promoting trail tourism.
Trail web map design must also consider whether alternative map formats, including
paper topographic maps and popular social fitness applications such as Strava, may be
more appropriate for trail users’ needs when actually on the trail. The ‘Map’ consideration
of the model proposed by Griffin et al. (2017) should involve not only adaptions in a specific
map but the provision of different solutions appropriate for different users and situations.
To enhance the value of participation for users using multiple solutions, integration
between application data, user contributions, and the participatory OpenStreetMap
project would be beneficial. OpenStreetMap data is often used in outdoor mobile appli-
cations. This integration could increase the value of participation to all users. Further research
will be required to align the needs of users, goals of trail managers and the OpenStreetMap
project, particularly in relation to spatial accuracy when the participatory model proposed by
Hauthal and Burghardt (2012) is fully implemented. OpenStreetMap is also not a suitable
location for storage of changing spatial information, comments or reviews though these
could be linked to OpenStreetMap objects representing physical trail features.
Feedback received in this research indicated the satisfaction of knowledgeable trail users
is determined more by the quality and quantity of information made available by a map
than by the ease of accessing this information. The more information is provided on a
map, the more knowledgeable trail users are inclined to contribute further information,
however the more utility is required to view information, the less usable the map
becomes for less knowledgeable users. Incorporating spatial participation into the UCD
methodology, even if through basic, paper-based methods, may be an effective method
of building this critical knowledge base and of gaining the support of local users. Knowl-
edge of specific users is a necessary prerequisite to creating map styles for particular
groups as was done by Dominguès and Hoarau (2011). This research suggests the provision
of both style and functionality adaptations would benefit various user groups, for example
specific modes for information contribution and a simplified view for new users.

Conclusion
This project demonstrated how modern open source technologies simplify the develop-
ment of database-driven single page web maps. Web 2.0 participatory applications in the
form of blogs, forums and social media are increasingly used by tourists to share infor-
mation about tourism opportunities. The potential of providing spatial participatory func-
tionality for the users of long-distance recreational trails has been recognised for some
time. This study provides a benchmark demonstrating the functionality for a specific trail.
Mapbox GL and React were used to create a web map allowing the display of information
requested by surveyed trail users and enabling these users to share both new spatial points
and further information related to existing features with other trail users. This application was
designed to allow a range of participatory utility through a highly usable interface. A scenario-
INTERNATIONAL JOURNAL OF CARTOGRAPHY 241

based survey was used to test the effectiveness and intuitiveness of specific features. This
evaluation indicated map design approaches were successful, but also identified a number
of areas for improvement. Many of the non-information-based feedback items required
only minor usability fixes, however one key area of missing functionality identified was fea-
tures allowing data moderation and expiry. A working participatory map application, moni-
tored over an extended period should be used to conduct further research into how
contributed spatial data can be effectively moderated and updated.

Acknowledgements
We wish to thank the Yorke Peninsula Council for their support of this project. We would particularly
like to acknowledge the support of Stephen Goldsworthy, Operations Manager. We would also like
to thank all survey participations for their input into the development and assessment of the web
map.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Joshua Gore is an Honours student at the University of South Australia, UniSA STEM, SA 5000,
Australia.
Stefan Peters is a Lecturer in GIS and Surveying at the University of South Australia, UniSA STEM, SA
5000, Australia. He focuses on applied geospatial science, remote sensing and cartography.
Delene Weber is an Associate Professor and Social Scientist at the University of South Australia,
UniSA STEM, SA 5000, Australia. He focusing on the human dimensions of environmental
management.

Data availability statement


A repository containing the website backend configuration and code is available at https://doi.org/
10.5281/zenodo.3526395. A repository containing the website frontend code is available at https://
doi.org/10.5281/zenodo.3526393.

ORCID
Joshua Gore http://orcid.org/0000-0002-1548-2474
Stefan Peters http://orcid.org/0000-0002-3604-4625
Delene Weber http://orcid.org/0000-0002-8293-4960

References
Confente, I. (2015). Twenty-five years of word-of-mouth studies: A critical review of tourism research.
International Journal of Tourism Research, 17(6), 613–624. https://doi.org/10.1002/jtr.2029
Dominguès, C., & Hoarau, C. (2011). On-demand cartography for trekkers. In A. Ruas (Ed.), Advances
in cartography and GIScience. Volume 1: Selection from ICC 2011, Paris (pp. 183–203). Springer
Berlin Heidelberg.
242 J. GORE ET AL.

Eftekhari, M. H., Barzegar, Z., & Isaai, M. T. (2011). Web 1.0 to web 3.0 evolution: Reviewing the impacts
on tourism development and opportunities (Vol. 6529, pp. 184–193). Springer, Berlin, Heidelberg.
Eugene, G. Y., Di, L., Rahman, M. S., Lin, L., Zhang, C., Hu, L., Shrestha, R., Kamg, L., & Yang, G. (2017).
Performance improvement on a web geospatial service for the remote sensing flood-induced crop
loss assessment web application using vector tiling. 2017 6th International Conference on Agro-
Geoinformatics, IEEE.
Griffin, A., White, T., Fish, C., Tomio, B., Huang, H., Sluter, C. R., Bravo, J. V. M., Fabrikant, S. I., Bleisch, S.,
Yamada, M., & Picanço, P. (2017). Designing across map use contexts: A research agenda.
International Journal of Cartography, 3(Supp. 1), 90–114. https://doi.org/10.1080/23729333.
2017.1315988
Hauthal, E., & Burghardt, D. (2012). Investigation and development of mobile touristic applications (1st
ed.). Springer.
Hotjar Ltd. (2019). Hotjar Analytics: Hotjar Ltd. https://www.hotjar.com/.
ISO. (2009). 9241-210: 2010. Ergonomics of human system interaction-part 210: Human-centred design
for interactive systems.
Kessler, F. (2011). Volunteered geographic information: A bicycling enthusiast perspective (report).
Cartography and Geographic Information Science, 38(3), 258. https://doi.org/10.1559/
15230406382258
Nivala, A.-M., Sarjakoski, T., Laakso, K., Itäranta, J., & Kettunen, P. (2009). User requirements for
location-based services to support hiking activities. Springer.
Nordan, R. P. V. (2012). An investigation of potential methods for topology preservation in interac-
tive vector tile map applications. In T. Midtbø (Ed.), NTNU, Institutt for bygg, anlegg og transport.
https://ntnuopen.ntnu.no/ntnu-xmlui/handle/11250/232116
Norman, D. A., & Draper, S. W. (1986). User centered system design: New perspectives on human-com-
puter interaction Hillsdale. Lawrence Erlbaum Associates.
Pack, S. (2019). Multimodal narratives case study on Facebook to promote outdoor tourism to mil-
lennials. In B. Johnson, S. Brautbar & A. Villegas (Eds.), ProQuest Dissertations Publishing. Grand
Canyon University.
Ricker, B., & Roth, R. (2018). Mobile maps and responsive design. Geographic Information Science &
Technology Body of Knowledge, 2nd Quarter 2018 Edition.
Roth, R., Ross, K., & Maceachren, A. (2015). User-centered design for interactive maps: A case study in
crime analysis. ISPRS International Journal of Geo-Information, 4(1), 262–301. https://doi.org/10.
3390/ijgi4010262
Roth, R., Young, S., Nestel, C., Sack, C., Davidson, B., Janicki, J., Knoppke-Wetzel, V., Ma, F., Mead, R., Rose,
C., & Zhang, G. (2018). Global landscapes: Teaching globalization through responsive mobile map
design. The Professional Geographer, 70(3), 395–411. https://doi.org/10.1080/00330124.2017.1416297
Sarjakoski, L. T., Kettunen, P., Flink, H.-M., Laakso, M., Rönneberg, M., & Sarjakoski, T. (2012). Analysis
of verbal route descriptions and landmarks for hiking. Personal and Ubiquitous Computing, 16(8),
1001–1011. https://doi.org/10.1007/s00779-011-0460-7
Sarjakoski, L. T., & Nivala, A. M. (2005). Adaptation to context – a way to improve the usability of mobile
maps. Berlin, Heidelberg: Springer.
Saura, J., Reyes-Menendez, A., & Palos-Sanchez, P. (2020). The digital tourism business: A systematic
review of essential digital marketing strategies and trends. In Digital Marketing Strategies for
Tourism, Hospitality, and Airline Industries. pp. 1–22. IGI Global.
Slocum, T. A., Cliburn, D. C., Feddema, J. J., & Miller, J. R. (2003). Evaluating the usability of a tool for
visualizing the uncertainty of the future global water balance. Cartography and Geographic
Information Science, 30(4), 299–317. https://doi.org/10.1559/152304003322606210
Tsou, M.-H. (2011). Revisiting web cartography in the United States: The rise of user-centered design.
Cartography and Geographic Information Science, 38(3), 250–257. https://doi.org/10.1559/
15230406382250
Yao, X., & Li, G. (2018). Big spatial vector data management: A review. Big Earth Data, 2(1), 108–129.
https://doi.org/10.1080/20964471.2018.1432115
Yoo, K.-H., & Gretzel, U. (2011). Influence of personality on travel-related consumer-generated media
creation. Computers in Human Behavior, 27(2), 609–621. https://doi.org/10.1016/j.chb.2010.05.002

You might also like