Content uploaded by Beat Signer
Author content
All content in this area was uploaded by Beat Signer on Nov 01, 2017
Content may be subject to copyright.
Bespoke Map Customization Behavior and Its Implications
for the Design of Multimedia Cartographic Tools
Anke M. Brock
University Toulouse
- ENAC
& Inria Bordeaux
Toulouse, France
anke.brock@enac.fr
Brent Hecht
PSA Computing
Research Group
Northwestern University
Evanston, IL, USA
bhecht@northwestern.edu
Beat Signer
Web & Information
Systems Engineering Lab
Vrije Universiteit Brussel
Brussels, Belgium
bsigner@vub.ac.be
Johannes Schöning
Human-Computer
Interaction
University of Bremen
Bremen, Germany
schoening@uni-bremen.de
ABSTRACT
While popular digital maps support an unprecedented
number of use cases, new reference map customization
tools have been created for purposes for which those maps
fall short. With the goal of informing the design of this new
class of cartographic tools, we present the first study of
naturalistic (“bespoke”) map customization behavior.
Through a mixed methods and mixed-media approach
involving a survey, the analysis of a corpus of customized
maps, and an interview with a power user, we find that
bespoke map customization is a relatively common activity
and identify frequent use cases as well as map
customization strategies. We discuss these use cases and
strategies in detail, and propose design implications for
future customization tools, such as the use of templates for
common use cases, adaptability for various customization
styles and the support of multimedia interaction.
Author Keywords
Digital maps; paper maps; geographic HCI (GeoHCI);
multimedia; cartography; design implications.
ACM Classification Keywords
H.5.m. Information interfaces and presentation (e.g. HCI):
Miscellaneous.
INTRODUCTION
While an age-old cartographic precept states that no single
map can be optimal for all purposes [27], popular digital
map platforms have made incredible strides towards
overcoming this barrier. Through innovations in map
interactivity and the use of intelligent algorithms, platforms
such as Google Maps, Apple Maps, and Bing Maps are now
capable of adapting to a large variety of use cases and
geographic areas.
The prevalence of a phenomenon that we call bespoke map
customization, however, suggests these mapping platforms
still do not adequately support certain use cases. In bespoke
map customization, end users leverage a variety of external
tools to transform a pre-existing map to better suit a
particular cartographic need. As we will discuss below,
common use cases for bespoke map customization include
Permission to make digital or hard copies of
all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. Copyrights
for compo
nents of this work owned by others than the author(s) must be
honored. Abstracting with credit is permitted. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior
specific permission and/or a fee. Request permissio
ns from
Permissions@acm.org.
MUM 2017, November 26–29, 2017, Stuttgart, Germany
© 2017 Copyright is held by the owner/author(s). Publication rights
licensed to ACM.
ACM ISBN 978
-1-4503-5378-6/17/11…$15.00
https://doi.org/10.1145/3152832.3152833
Figure 1: Excerpt of a travel map indicating routes
between different points of interest, customized with
MS Paint (map corpus, P6).
constructing maps for planning vacations, inviting visitors
to an upcoming event (e.g. a wedding), and supporting the
spatial referencing of memories.
We define bespoke map customization more specifically as
the act of producing a new, simple spatial dataset with a
small number of features and simultaneously visualizing
that dataset without the use of programming. As we will
discuss below, the mechanisms employed in bespoke map
customization are highly diverse and range from the use of
MS Paint to complex workflows that span over different
media and involve printing maps, annotating them, and then
scanning them back into their computer.
To our knowledge, no prior research has been conducted on
bespoke map customization behaviors, practices, and use
cases. In this paper, we address this gap in the literature,
focusing specifically on the most commonly used map type:
reference maps (i.e. maps used for navigation and orien-
tation). To gain a broad perspective on bespoke map
customization, we adopted a mixed methods and mixed-
media approach. Specifically, our study includes an online
survey, qualitative analyses of a corpus of customized
maps, as well as an interview with a power user.
Our results suggest that bespoke map customization is a
common activity—over 70% of the survey respondents
reported having customized a map—and provide support
for a series of design implications. For instance, our results
reveal the importance of designing for mixed-media map
support, e.g. treating the possibility to print customized
maps as a first-tier feature and supporting the digitization of
paper-based annotation done in-situ in a mobile context
while engaging in an activity. Moreover, our results reveal
the most common use cases for map customization and
highlight the importance of customized maps as memory
aids and mementos following the completion of an activity.
The importance of our study has increased with the growing
number of tools that allow web users to make personalized
maps without programming. The most popular tool is
Google My Maps (https://www.google.com/mymaps).
Others include ZeeMaps (http://zeemaps.com) and Map-
maker (http://mapmaker.education.nationalgeographic.com)
by National Geographic. The design implications from our
work are largely applicable to these tools. Many of the
implications suggest significant new features for the entire
class of existing customization tools, and other implications
support the design decisions of some tools over others’.
To summarize, the contributions of this paper are as
follows: Using a mixed methods and mixed-media
approach, we present the first exploration of bespoke map
customization behavior. We suggest a series of design
implications for reference map customization tools such as
templates for common use cases (e.g. travel or meeting
maps), adaptation to various customization styles (e.g. free-
hand annotation) and support of multimedia interaction.
RELATED WORK
This paper draws motivation from four research streams in
the literature: annotation in information visualization,
spatial decision support systems, existing map
customization technologies, and the comparison of paper
and digital media for cartography. In the following we
discuss each stream in turn.
Annotation in Information Visualization
Reference map customization activities can be considered a
subset of annotation in information visualization, although a
subset with a number of unique properties. A major focus in
recent annotation work has been on asynchronous
collaboration around visualization of annotations [4,11] and
on the automatic generation of annotations that highlight
comparisons made in corresponding text [9,13]. Reference
map customization is distinguished from this literature in
that it involves maps that are intended for navigation and
orientation (the broad use cases for reference maps), rather
than thematic maps, which are intended to communicate the
geographic distribution of an attribute (e.g. population, or
temperature) [29]. Annotations in thematic maps are
targeted at highlighting trends rather than supporting
navigation, orientation and the diverse use cases of
reference maps outlined below.
Spatial Decision Support Systems
Map customization is also related to spatial decision
support systems, which have been heavily studied (e.g. [1]).
These systems help decision makers such as disaster relief
specialists to quickly analyze scenarios, often by using
visualization tools. Bespoke reference map customization is
distinct from these systems as it addresses casual users,
whereas spatial decision support systems are for pro-
fessionals and experts. This results in highly divergent use
cases and design challenges (as we will see below).
Map building tools for geowikis such as OpenStreetMap
(OSM, http://www.openstreetmap.org) present an offshoot
of spatial decision support systems that are related to map
customization. These tools have been in the focus of recent
research at the interface of Geography, Multimedia and
HCI (e.g. [10,26]) and the volunteered geographic infor-
mation research community more broadly (e.g. [7,16,23]).
However, like their more expert-oriented analogues, these
tools can have different use cases (i.e. disaster mapping)
than reference map customization. Geowikis like OSM bear
some similarities to reference map customization tools, but
their goal is to create a single, uniform reference map rather
than to help users in customizing their maps for their own
individual use cases (i.e. map creation is different from map
customization).
Existing Map Customization Technologies
A major motivation for this work comes from the recent
increase in available reference map customization tools. By
far the most popular reference map customization tool is
Google My Maps, which was launched in 2007 and has
been through a number of iterations since then. Google
My Maps is a general reference map customization tool that
supports a number of features such as adding POI “pins”
and routes to a Google basemap. National Geographic
Mapmaker Interactive and ZeeMaps are competitors, but
the latter has a more professional focus (e.g. including
specialized tools for journalists), while the former is
focused on educational audiences. Tools such as MapBox
Studio (https://www.mapbox.com/mapbox-studio/) are also
related to map customization. However, these tools require
programming skills as they make use of JSON-based
styling specifications and thus do not qualify as map
customization as defined above. The leading GIS company
ESRI also targets this general space with its StoryMaps
(http://storymaps.arcgis.com), but like the above-mentioned
tools, they require programming skills.
Map customization is also related to the notion of the “map
mashup”, in which pre-existing datasets are added to an
online map through APIs or technologies like Google
Fusion Tables. However, mashups often require some
programming skills (or at least knowledge of related topics
like database joins) and mashups are almost always
thematic maps and not reference maps.
Paper and digital media for cartography
As reported below, people use mixed-media techniques
(paper and digital) in bespoke map customization. This is
related to the literature on the comparative advantages of
paper and digital media for cartographic tasks. While the
dynamic displays that digital maps afford are incredibly
powerful, paper maps retain certain advantages. For
instance, paper maps do not require power; their light
weight can make them more portable than digital maps on
mobile devices [22], their large display sizes relative to
mobile maps support certain activities better [5,15], such as
collaboration [2]; they can be cheaper when engaging in
foreign travel due to roaming costs [17]; and they can be
easier to customize while in the field [22]. Hurst et al. [14]
showed that geographic knowledge and the context of use
influenced whether users preferred paper or digital media:
as geographic skills increase, there is a greater preference
for paper maps, while non-experts prefer digital maps.
METHODOLOGY
The primary goal of our research was to understand the
design space for bespoke map customization tools and
provide design implications for the design of existing and
future tools. In order to obtain sufficiently diverse per-
spectives on map customization activities with respect to
this goal, we applied a mixed methods approach involving
three separate studies: (1) an online survey, (2) the analysis
of a corpus of customized maps, and (3) an interview with a
power user. With the survey and analysis of the map
corpus, we used an open, bottom-up approach to understand
which factors are important. The interview with a power
user afforded us a deeper understanding of those map
customization factors with Google My Maps, one of the
most commonly used map customization tools.
Online Survey
To design our survey, we drew methodological motivation
from Morris et al.’s influential exploratory work on social
media question asking (SMQA) [20]. Morris and colleagues
performed the first exploration of SMQA through the
deployment of a snowball-sample survey. Thanks to the
novelty of the SMQA research space at the time, their work
has proven to be significant, having identified a number of
key characteristics of SMQA behavior that have led
researchers to engage in more targeted and robust analyses.
Reference map customization in 2017 is in the same
position as SMQA in 2009: practiced by users, but not well
understood at a high level by researchers. As such, we
designed our survey in a similar fashion as [20], trading off
robust sampling strategies (which are difficult to define for
this novel space) for a practical exploration of a design
space and high-level insights, with the aim of supporting
more targeted follow-up work.
A link to our online survey was distributed internationally
to a mostly non-expert population with respect to mapping
tools via professional and student computer science and
human-computer interaction mailing lists as well as through
social media platforms in our personal and professional
network in 2014. We explicitly invited people who custom-
ized maps and people who did not customize maps to par-
ticipate in the survey. The survey was conducted in English
using LimeSurvey (https://www.limesurvey.org). In total,
we received 108 completed responses (39 female, 69 male;
average age 35.2 years, age range: 21-61, SD: 7.61).
The order and type of questions in the questionnaire
depended on participants’ prior experience with map
customization. The survey first asked participants whether
they had previously engaged in any bespoke map
customizations. In the case of a negative reply, the survey
presented a single question inquiring the reasons. On the
other hand, if respondents had engaged in map
Did you customize digital maps, paper maps or both of them
and how many digital or paper maps did you customize?
(estimate)
What were the reasons for customizing these maps?
Please specify the type of the online map service you used.
For which purpose did you customize the map?
In general, when have you annotated the map relative to the
activity you used it for?
Describe what you have customized.
How have you annotated the map? (using which tools)
Have you made the annotation for yourself or to communicate
the map to another person?
Have you used (are you going to use) the map multiple times?
Did you share your map/annotations with somebody else?
If available, please upload a picture of the customized map(s).
Table 1: Survey questions as presented to participants who
had previously customized maps.
customization, they were asked about the character and
frequency of their map customization activities (see
Table 1). The average time to complete the survey with all
questions was around 20 minutes. The gathered data was
quantitative (7-point Likert scale) and qualitative (open-
ended questions). The quantitative data was analyzed using
descriptive statistics; the qualitative data was analyzed
using a bottom-up coding scheme as defined below.
CORPUS ANALYSIS
We asked survey respondents to submit screenshots, scans,
and photos of their maps if available. We collected a total
of 38 maps from 17 different users (10 male, 7 female;
average age 35.7 years, range: 26-58, SD: 9.71). Table 2
presents details about participants and maps. Our final
corpus consisted of 37 maps from 16 users. We removed
one map that was a map mashup (as defined above) and not
a customized map. Our dataset consisted of maps from
12 countries on four continents, with the most maps coming
from Germany (8), followed by France (7) and the United
States (6). A subset of those maps is depicted in this paper
and the maps considered here are available for research
purposes upon request.
Our analysis of the corpus utilized a map coding procedure
closely following the one proposed by Schöning et al. [27]:
two coders started with a single pass through the map using
the coding dimensions defined in [27] and identified
dimensions to be used in the coding exercise and others that
did not apply. Next, the two coders made a single pass on
the data set to identify additional dimensions of analysis.
Then, they merged their coding scheme. This process
resulted in a total of 21 dimensions, a subset of which are
described in Table 3. Once this final scheme was estab-
lished, each map was coded independently by two coders.
Then, a third independent coder resolved disagreements
(only four cases in total). As recommend in [27], we as-
sured that at least one of the coders was trained in cartog-
raphy (at least two university-level cartography courses).
INTERVIEW WITH A POWER USER
Complementary to the survey with casual users, we wanted
to learn more about how map customization is done by a
“power” user with one specific tool and whether there are
similarities or differences with casual users. We invited a
woman (29 years) from our personal network who custom-
izes maps more often than average people in their daily
lives, but not professionally. As the customization tool we
selected Google My Maps as it is among the most
commonly used map customization tools. At the time of the
interview, the power user had been using Google My Maps
for more than six years and had created over 20 maps with
Google My Maps. Her maps covered a broad range of use
cases, including travel, exploring new neighborhoods or
creating maps for her blog. In a one-to-one session that
lasted for around an hour and a half, the interviewee walked
the interviewer through the different maps she had created,
describing for each map when, how and why it has been
created, and if the map had been shared with others. This
interview also allowed us to access additional examples of
customized maps.
Table 2: Characteristics of participants and maps from the map corpus.
Table 3: Overall coding scheme for the map corpus.
1.
Figure 3. Using various marker types to highlight POIs
on a travel map (map corpus, P8). Google Maps was used
as base map ©Google 2014.
Figure 4: Example of a meeting map with scribbled anno-
tations using MS Paint (map corpus, P11).
Figure 5: Meeting map, based on local public transport-
ation map customized with MS Paint (P12).
RESULTS
We report our findings from all three studies together,
structured by main themes that we identified in our data
sets.
#Theme 1: Many People Customize Maps, but Most
Remain Novices
Like for Morris et al.’s study [20], our survey design allows
to give a broad overview of the popularity of bespoke map
customization, but not a precise percentage of the
population that has engaged in the activity. Yet, the broad
picture in our survey results regarding the prevalence of
map customization was a clear one: bespoke map custom-
ization is far from a niche activity and has been engaged in
at least once by a non-trivial swath of the population. More
specifically, we found that the majority of the survey
participants (71%) had customized at least one reference
map. Even adjusting for some bias in recruitment, this is a
substantial percentage for a bespoke technological en-
deavor.
When asked which maps they had customized in the past
year, 62 participants (57%) reported having customized
online maps, 50 (46%) paper maps, and 33 (36%) online as
well as paper maps. When asked for the number of maps
customized in the past year, the mean number of digital
maps was 6.4 (SD: 5.9) and for paper maps 4.3 (SD 3.7).
Among the 29% of participants who had never done any
map customization, the majority (57%) reported not feeling
the need to do so, 12% did not want to “destroy the map”
by writing on it, and 10% did not know it was possible.
With 71% of respondents reporting some map custom-
ization activity, these results substantiate the importance of
designing for bespoke map customization. However, the
results also reveal that, on average, people customize a
rather small number of maps per year (6.4 digital and 4.3
paper maps). We can thus assume that map customization is
generally not a task at which users become experts.
Consequently, designers need to be careful to design map
customization systems for primarily novice, inexperienced
users.
#Theme 2: Map Customization Has Several Primary Use
Cases
We were able to identify several common primary use cases
for map customization (see Figure 2). In the map corpus
coding schema (Table 3), the “purpose” category served for
identifying the use cases. In the survey and interview, we
specifically asked participants for which purposes they
customized maps.
The most important use case in all three sources was
tourism. This was particularly apparent in our map corpus,
Figure 2. Most common use cases for map customization for
digital maps (dotted blue) and paper maps (orange diagonal
stripes). The category “other” includes different uses cases
that were each named only once. We do not further discuss
“wayfinding” as this is a basic purpose of reference maps
and also included in some of the other categories.
Figure 6: Excerpt of a customized map serving multiple purposes: 1) a wedding invitation, indicating venues with different
customized markers, 2) an artifact in a photo album after the event (P8).
2.
in which 20 maps (54%) were customized to support
touristic activities (e.g. Figures 1, 3 and 9). In the survey,
17 participants reported customizing maps for traveling and
8 maps by our power user interviewee had been customized
for this purpose. Through the different sources, we could
identify that bespoke maps for tourism mainly served to
highlight POIs and to support wayfinding (e.g. visualization
of travel routes and taking maps along for navigation). The
POIs that participants added to maps included their accom-
modation, landmarks, shops, restaurants, airports and public
transportation. Three people added street addresses and
phone numbers to their maps.
The second most-common use case for bespoke map
customization was spatially-referenced memories in photo
albums, diaries or on blogs. A total of 9 survey participants
mentioned that they used maps for this purpose. Par-
ticipants explained that these maps served as mementos for
places and locations they had visited and the routes that
they took. Within the map corpus, we identified three maps
that solely served this purpose. These maps depicted
general areas of cities and contained around 20 markers on
average. Interestingly, in addition to these three maps, most
of the travel maps were used as a memento after the trip.
Memory maps also played an important role for our
interviewee who has a travel blog and had customized
seven maps for this purpose.
Another prominent category in our map corpus was
“meeting maps”, with a total of seven maps. Meeting maps
usually depicted a city and contained a single marker
highlighting a specific place as well as directions indicating
how to get there from various origins (e.g. text or arrows,
see Figures 4 and 5). Five survey participants mentioned
that they used bespoke customized maps to explain routes
to friends or family and to indicate meeting points. More-
over, our interviewee created one map with the specific
purpose of explaining her visitors how to reach her flat
from the airport. Very similar to meeting maps are “event
maps”, which highlight venues that are important for an
event, e.g. a wedding or a conference (see Figure 6). The
map corpus contained five of these maps and five survey
participants mentioned having created them.
The maps in our study also played an important role in
supporting sports and leisure activities, including hiking,
cycling, cross-country running, hot air ballooning and
flying a plane. The map corpus contained four maps of this
type, and eight survey participants reported that they used
maps for leisure and sports. One survey participant reported
that for hiking he annotated maps with lift schedules,
potential alternatives routes, the time required to walk along
these alternatives, as well as picnic areas next to the trail.
These annotations made him feel safer when hiking as he
had a variety of options in case he had to diverge from the
initial plan.
Fourteen survey participants reported that they developed
bespoke maps to help orient themselves in a new geo-
graphic environment. Our interviewee created three maps
when she moved to a new city to support the familiarization
process with the environment.
In all of the above-mentioned use cases, bespoke
customized maps for the same use case tend to be quite
similar in character. This suggests that designers may be
able to better support these use cases by providing tem-
plates which users could select based on their needs. For
instance, if the user selects to create a meeting map, a
template-based system could propose a marker tool for
highlighting the meeting point as well as the possibility to
add textual descriptions and graphical directions (e.g. arrow
symbols) to the map. Similarly, customized travel maps
typically contain a specific set of POIs such as
accommodation and key public transport locations, as well
as routes between these and other POIs. A basic version of
the template-based approach we suggest has been
implemented for at least one use case: wedding maps
(weddingmapper.com). Our results suggest that this
approach would likely be effective for most of the popular
use cases of bespoke map customization.
It is straightforward to imagine tools that begin creating
these maps automatically by mining interesting informa-
tion. This can be thought of as the spatial equivalent of
systems that mine e-mail for hotel reservations and
automatically add them to associated calendar systems
(e.g. Gmail and Google Calendar). Also, if a certain user
generally annotates specific points of interests (e.g. art mu-
seums, public transport or Thai restaurants), the system
could automatically annotate the map, suggesting points of
interests that match previous annotations using algorithms
from machine learning and artificial intelligence. Also, it
has been proposed to automatically detect points of interests
in geo-referenced photos [30], videos [31] or social media
[25]. To our knowledge, no existing map annotation tool so
far makes use of machine learning algorithms or artificial
intelligence.
#Theme 3: Maps are Customized Throughout an
Activity's Lifecycle
Our findings suggest that bespoke map customization can
take place before, during and/or after the activity to which
the map relates. Indeed, our survey results indicate that
around 57% of maps were customized before the activity,
about 24% of maps during the activity, and 19% of maps
after the activity (Figure 7).
Although most maps were customized before the activity,
“memory maps” presented an important use case in which
customizing maps after the activity needs to be supported.
Furthermore, the “travel maps” use case suggests the need
to customize maps also during the activity. One survey
participant reported that for traveling, he used digital maps
created with Osmand (http://osmand.net/), an OSM- and
Wikipedia-based map service, which can be used offline on
a mobile Android device. This allows him to add POIs both
prior to departure and while traveling without any roaming
fees for accessing the Internet abroad.
In accordance with our findings, we suggest that map
customization tools should provide the possibility to
customize the same map during different stages of a map’s
lifecycle. As devices used for accessing the maps may vary,
the map customization tools need to support different
devices, media and interaction techniques (paper maps and
pen; desktop computer with large screen, keyboard and
mouse; smartphone, tablet or other mobile devices with
smaller screen and touch display). The possibility to access
Google My Maps from a smartphone is quite recent and we
believe that further progress could be achieved regarding
this interaction scenario (e.g. not all functionalities that
Google My Maps provides on a computer are also available
on the application for mobile devices).
The use of customized maps on mobile devices might also
additionally provide the possibility to highlight a user’s
current position or log a user’s track as already proposed by
some mobile applications, but not yet included in map
customization tools. This could make it easier to annotate
maps in one’s surroundings, and facilitate the navigation to
places that have previously been annotated.
Beyond that, map customization tools need to support
offline map annotation—similar to Osmand as described
above—for supporting customization of maps during travel
in a foreign country where accessing online maps and
services is not affordable due to roaming. We see that this
feature is currently not well supported by common map
customization tools.
#Theme 4: Map Customization Involves Hybrid Use of
Different Media (Paper and Digital)
Among the survey participants who had previously
customized maps, 79% used online maps and 65% used
paper maps. Almost half of them (47%) used both. With
regard to the latter, a number of participants reported
following a hybrid approach in which the same map was
customized digitally and on paper. As an example, some
users digitally customized a map, then printed it and further
annotated the printed map during a city trip. Indeed, while
the most popular map customization strategy was to directly
annotate on Google My Maps (41%, see Figure 8), the
second most popular map customization strategy involved
generating a desired map excerpt using a digital map,
printing the map, and customizing the printout using a pen
(13%). While 12% of all paper maps were digitized at a
later stage, one participant applied the opposite process as
he traced the original map by hand, scanned it and then
worked with it on digital media. We thus observed that map
customization is often a hybrid process: people mix analog
and digital spatial information and tools.
Interestingly, we observed that only 20% of all survey
participants who had customized paper maps had paid for
Figure 7: When survey participants customized papers
relative to the activity (marked in vertical blue patterns for
paper and in horizontal orange patterns for digital maps).
Figure 8: Most common customization techniques.
“Other” includes 19 different techniques that were each
named only once.
these maps. Free maps—e.g. from tourist offices—were
used by 38% and 40% had printed free maps from the
Internet. One survey participant explicitly stated that he did
not customize the paper maps included in his travel guide
books because it is “bad to write in books”. As mentioned
above, 12% of the respondents who had never customized
maps, stated that they did not want to destroy maps by
writing on them.
Our results suggest that future reference map customization
tools should not solely focus on the digital production and
use of maps, but should also support the integration of
paper into the customization workflow. To fully support the
types of mixed-media bespoke activities that we observed,
this integration needs to go substantially beyond the surface
level. As an example, on digital maps, labels can be shown
dynamically depending on the user input (e.g. mouse/touch
position and movement). However, on a static paper map,
significant care must be taken in optimizing label place-
ment. A similar issue exists for the map legend: when
zooming and panning is not possible as is the case for paper
maps, the legend must be placed in the correct location to
avoid occluding key parts of the map [6]. Implementing
these features, e.g. through a process prior to printing a
map, will likely require some combination of automated
approaches (label placement is a challenging algorithmic
problem) as well as provision for user customization. Most
importantly, the user should not be expected to be an expert
in the effects of converting a map from digital to print
format (e.g. font size issues, legend occlusion) prior to
printing. The possibilities should be highlighted for the user
a priori and solutions should be proposed. Knowledge from
fields such as machine learning, artificial intelligence and
image recognition could provide significant improvements
to this regard.
The need for mixed-media support also opens up
opportunities for future work in the fields of HCI and
augmented reality. For instance, it has previously been
studied how to augment paper maps by using Anoto’s
digital pen technology [21,24,28], or how to project digital
maps which can then be annotated using regular pens [3].
Recently, Fedosov et al. have designed a wearable aug-
mented reality system that allows skiers to share content on
a printed panoramic resort map using a head-mounted
display [8]. However, so far most of these systems remain
in the laboratory and have not been transferred to the real
world.
#Theme 5: Map Customization Tools Should Support
Diverse Cartographic Styles
The results of all three data collection methods reveal that,
while there are high-level similarities in the cartographic
styles of bespoke maps, there is a great deal of diversity
with respect to lower-level stylistic decisions. For instance,
in the case of the use of map markers, our results suggest
that the highlighting of POIs on maps by using markers
(e.g. pins, crosses or circles) is an important activity in
bespoke map customization. Indeed, almost all maps in our
corpus contained some form of markers. Highlighting
points with markers was also the most popular answer to
the survey question “Describe what you have customized”:
31% of all answers regarding the online map and 35% of
the answers regarding the paper map referred to this,
followed by customization of other map elements such as
paths, adding a legend, or manipulating map perspective
and orientation.
However, while adding markers is a common bespoke map
customization activity, that is where the similarities end.
People used a wide variety of marker types, used markers
for diverse purposes, and added markers using a diverse set
of tools. For instance, our interview participant heavily
customized markers and, using Google My Maps,
developed a color code to differentiate between “must see”,
“optional” and “leave out” destinations or different
categories (Figure 9). Figure 3 shows a map in which
different predefined markers have been used. When
annotating by hand, people frequently scribbled markers
with various, informal shapes. The same behavior was
observed when people annotated maps using MS Paint
(Figures 1, 4 and 5), with 7% of survey respondents using
Paint as a map customization tool. Existing map
customization tools could learn from personalization of
menus and options which has been successfully employed
in other domains [18].
Figure 9: Travel map created with Google My Maps:
modified basemap makes customized POIs more prominent;
use of color code for markers to distinguish different
categories of sites (power user interview). Google Maps was
used as base map ©Google 2014.
Figure 10: Using “fun” marker (smileys) and arranging them
to form a larger smiley, map created with Bing Maps (P8).
©Microsoft 2014.
Moreover, we saw examples of participants who employed
markers for purposes well outside the typical reference map
considerations of navigation and orientation. In particular,
we saw evidence of the cooption of traditional map markers
for ludic purposes (i.e. “fun”). This can be seen in Figure 10
where, in addition to utilitarian markers, we see that the
customizer has also arranged some markers into a “smiley
face”. To reinforce the ludic aspects of the map, some of
the markers were themselves smileys. Existing map cus-
tomization tools (and mapping platforms) do not support
ludic activities like these. While our study suggests that the
substantial majority of bespoke maps are primarily utilitar-
ian in nature, providing a set of ludic features may be desir-
able. For instance, the use of full emoji character sets (pic-
tographs that are popular in text-based communication [19])
as map markers rather than simple smiley faces is one
simple way to work towards this.
In addition to the diverse employment of markers, the
addition of diverse labels is another important cartographic
activity in bespoke map customization. Around 20% of the
maps in the corpus contained bespoke labels. These labels
were often very informal in nature (e.g. scribbles and
sketches), and primarily occurred with maps generated
using MS Paint and on printed paper maps annotated using
pen and pencil (Figures 1, 3 and 5). While this has been
proposed in different contexts, such as for free-hand anno-
tation of endoscopic videos [12], this type of scribbling is
not widely supported in map customization tools, even as
pen input becomes increasingly democratized (e.g. via
Microsoft Surface devices and Apple iPads). Our results
suggest that simply adding pen input capability when a
hardware device supports it could go a long way towards
better facilitating naturalistic map customization activities.
DISCUSSION AND DESIGN IMPLICATIONS
The purpose of this paper is to provide insight into diverse
bespoke map customization activities to support the design
of better cartographic tools. To summarize, we propose the
following design implications for current and future map
customization tools:
Provide templates and tools for different use cases:
We observed that map customization has several primary
use cases, and each has relatively well-defined cartographic
needs. As such, it would be possible to build templates into
cartographic tools to better facilitate the creation of specific
customized maps.
Increase the flexibility and personalization of such tools:
Map customization tools should support various
cartographic styles, e.g. handwriting and scribbling, or a
large choice of marker types. Support for ludic activities is
also suggested as a new feature.
Support map customization at different moments:
There is no single fixed time in the course of an activity for
the creation of bespoke customized maps. Thus, map cus-
tomization tools need to provide the possibility to custom-
ize maps before, during and after an activity. Therefore,
tools need to cater for different devices (e.g., computer,
smartphone, paper). Online and offline use should be
supported even on mobile devices, as well as the possibility
to upload offline annotations later.
Support multimedia interaction:
People go back and forth between different media when
customizing maps. Digital maps should be convertible into
paper maps using printing options that go beyond those that
are currently available (e.g. adapting legend style and place-
ment or the amount of displayed text). Advanced interactive
devices such as augmented reality systems or Anoto pens
could be used to make paper maps interactive.
Make tools more intelligent:
We identified opportunities for the integration of machine
learning approaches into map customization tools. In
particular, information extraction approaches like those
taken in modern calendaring systems could be applied to
automatically-generated travel and meeting maps, a
common use case for bespoke customization.
While we presented the best available data to date about
bespoke map customization frequency and behavior, our
results have some limitations. First, out of the 37 maps in
the map corpus, only three maps were created on paper. We
suppose that there was a bias towards submitting digital
maps given that the maps were collected through an online
survey. In the future, it would be interesting to have a closer
look at how paper map customization is done in order to
design multimedia tools. Additionally, while our survey
sampling approach was appropriate for gaining a first-pass
understanding of map customization behavior, a more
traditional survey framed by the findings presented here is
an obvious next step for further research in this domain.
CONCLUSION
In this paper, we presented the first exploration of the
common activity of bespoke map customization. We
applied a mixed methods approach involving a survey, an
interview with an expert, and the analysis of a corpus of
customized reference maps (which we are making available
to the public upon request). We also took first steps towards
systematically deriving design implications for current and
future map customization tools to better support the use
cases for which bespoke map customization is employed.
Our design implications suggest a series of new features for
mapping tools, such as templates for common custom-
ization use cases, more personalization options and deep
support for multimedia (paper and digital) map annotation.
ACKNOWLEDGMENTS
We thank the survey participants and the power user
without whom this work would not have been possible.
Furthermore, we want to thank our colleagues who helped
analyzing the map corpus.
REFERENCES
1. G. Andrienko, N. Andrienko, P. Jankowski, et al. 2007.
Geovisual Analytics for Spatial Decision Support:
Setting the Research Agenda. International Journal of
Geographical Information Science 21, 8: 839–857.
http://doi.org/10.1080/13658810701349011
2. Barry Brown and Matthew Chalmers. 2003. Tourism
and Mobile Technology. Proceedings of the European
Conference on Computer Supported Cooperative Work
(ECSCW 2003), Springer Netherlands, 335–354.
http://doi.org/10.1007/978-94-010-0068-0_18
3. Julia Chatain, Marie Demangeat, Anke Brock, Didier
Laval, and Martin Hachet. 2015. Exploring Input
Modalities for Interacting with Augmented Paper
Maps. 27ème conférence francophone sur l’Interaction
Homme-Machine (IHM 2015), ACM.
http://doi.org/10.1145/2820619.2825002
4. Mauro Cherubini. 2008. Annotations of Maps in
Collaborative Work at a Distance.
http://doi.org/10.5075/epfl-thesis-4116
5. J. Doyle, M. Bertolotto, and D. Wilson. 2009.
Evaluating the Benefits of Multimodal Interface
Design for CoMPASS: A Mobile GIS. GeoInformatica
14, 2: 135–162.
http://doi.org/10.1007/s10707-009-0079-2
6. Jason Dykes, Jo Wood, and Aidan Slingsby. 2010.
Rethinking Map Legends with Visualization. IEEE
Transactions on Visualization and Computer Graphics
16, 6: 890–899. http://doi.org/10.1109/TVCG.2010.191
7. Thore Fechner, Dennis Wilhelm, and Christian Kray.
2015. Ethermap: Real-time Collaborative Map Editing.
Proceedings of the ACM Conference on Human
Factors in Computing Systems (CHI 2015), ACM
Press, 3583–3592.
http://doi.org/10.1145/2702123.2702536
8. Anton Fedosov, Evangelos Niforatos, Ivan Elhart,
Teseo Schneider, Dmitry Anisimov, and Marc
Langheinrich. 2016. Design and Evaluation of a
Wearable AR System for Sharing Personalized Content
on Ski Resort Maps. Proceedings of the International
Conference on Mobile and Ubiquitous Multimedia
(MUM 2016), ACM Press, 141–152.
http://doi.org/10.1145/3012709.3012721
9. Tong Gao, Jessica R. Hullman, Eytan Adar, Brent
Hecht, and Nicholas Diakopoulos. 2014. NewsViews:
An Automated Pipeline for Creating Custom
Geovisualizations for News. Proceedings of the ACM
Conference on Human Factors in Computing Systems
(CHI 2014), ACM Press, 3005–3014.
http://doi.org/10.1145/2556288.2557228
10. Brent Hecht, Johannes Schöning, Muki Haklay, et al.
2013. Geographic Human-Computer Interaction.
Extended Abstracts on Human Factors in Computing
Systems (CHI EA 2013), ACM Press, 3163–3166.
http://doi.org/10.1145/2468356.2479637
11. Jeffrey Heer, Fernanda B. Viégas, and Martin
Wattenberg. 2007. Voyagers and Voyeurs: Supporting
Asynchronous Collaborative Information
Visualization. Proceedings of the ACM Conference on
Human Factors in Computing Systems (CHI 2007),
ACM Press, 1029–1038.
http://doi.org/10.1145/1240624.1240781
12. Marco A. Hudelist, Sabrina Kletz, and Klaus
Schoeffmann. 2016. A Tablet Annotation Tool for
Endoscopic Videos. Proceedings of the ACM
Multimedia Conference (MM 2016), ACM Press, 725–
727. http://doi.org/10.1145/2964284.2973822
13. Jessica Hullman, Nicholas Diakopoulos, and Eytan
Adar. 2013. Contextifier: Automatic Generation of
Annotated Stock Visualizations. Proceedings of the
ACM Conference on Human Factors in Computing
Systems (CHI 2013), ACM Press, 2707–2716.
http://doi.org/10.1145/2470654.2481374
14. P. Hurst and P. Clough. 2013. Will we be Lost Without
Paper Maps in the Digital Age? Journal of Information
Science 39, 1: 48–60.
http://doi.org/10.1177/0165551512470043
15. Hans-Christian Jetter, Svenja Leifert, Jens Gerken,
Sören Schubert, and Harald Reiterer. 2012. Does
(Multi-)Touch Aid Users’ Spatial Memory and
Navigation in “Panning” and in “Zooming & Panning”
UIs? Proceedings of the International Working
Conference on Advanced Visual Interfaces
(AVI 2012), ACM Press, 83–90.
http://doi.org/10.1145/2254556.2254575
16. Marina Kogan, Jennings Anderson, Leysia Palen,
Kenneth M. Anderson, and Robert Soden. 2016.
Finding the Way to OSM Mapping Practices:
Bounding Large Crisis Datasets for Qualitative
Investigation. Proceedings of the ACM Conference on
Human Factors in Computing Systems (CHI 2016),
ACM Press, 2783–2795.
http://doi.org/10.1145/2858036.2858371
17. Colin F. Mang, Linda A. Piper, and Natalya R. Brown.
2016. The Incidence of Smartphone Usage Among
Tourists. International Journal of Tourism Research 18,
6: 591–601. http://doi.org/10.1002/jtr.2076
18. Joanna McGrenere, Ronald M. Baecker, and Kellogg
S. Booth. 2002. An Evaluation of a Multiple Interface
Design Solution for Bloated Software. Proceedings of
the ACM Conference on Human Factors in Computing
Systems (CHI 2002), ACM Press, 164.
http://doi.org/10.1145/503376.503406
19. Hannah Miller, Jacob Thebault-Spieker, Shuo Chang,
Isaac Johnson, Loren Terveen, and Brent Hecht. 2016.
“Blissfully Happy” or “Ready to Fight”: Varying
Interpretations of Emoji. International AAAI
Conference on Web and Social Media (ICWSM 2016),
259–268.
20. Meredith Ringel Morris, Jaime Teevan, and Katrina
Panovich. 2010. What Do People Ask Their Social
Networks, and Why?: A Survey Study of Status
Message Q&A Behavior. Proceedings of the ACM
Conference on Human Factors in Computing Systems
(CHI 2010), ACM Press, 1739–1748.
http://doi.org/10.1145/1753326.1753587
21. Moira Norrie and Beat Signer. 2005. Overlaying Paper
Maps with Digital Information Services for Tourists.
Proceedings of the International Conference on
Information Technology and Travel & Tourism
(ENTER 2005), Springer Vienna, 23–33.
https://doi.org/10.1007/3-211-27283-6_3
22. Volker Paelke and Monika Sester. 2007. Design
Exploration of Augmented Paper Maps. In Proceedings
of the ISPRS Workshop Visualization and Exploration
of Geospatial Data: 7.
23. Leysia Palen, Robert Soden, T. Jennings Anderson, and
Mario Barrenechea. 2015. Success & Scale in a Data-
Producing Organization: The Socio-Technical
Evolution of OpenStreetMap in Response to
Humanitarian Events. Proceedings of the ACM
Conference on Human Factors in Computing Systems
(CHI 2015), ACM Press, 4113–4122.
http://doi.org/10.1145/2702123.2702294
24. Christian Pauschert, Emanuel Riplinger, Carola Tiede,
and Volker Coors. 2011. Benefits Through Linking of
Analogue and Digital Maps. In Advances in
Cartography and GIScience. Volume 1 (Lecture No),
Anne Ruas (ed.). Springer Berlin Heidelberg, 205–217.
http://doi.org/10.1007/978-3-642-19143-5_12
25. Adrian Popescu and Aymen Shabou. 2013. Towards
Precise POI Localization With Social Media.
Proceedings of the ACM International Conference on
Multimedia (MM 2013), ACM Press, 573–576.
http://doi.org/10.1145/2502081.2502151
26. Rossano Schifanella and Bart Thomée. 2016. The
Lifecycle of Geotagged Multimedia Data. Proceedings
of the ACM on Multimedia Conference (MM 2016),
ACM Press, 1471–1472.
http://doi.org/10.1145/2964284.2986911
27. Johannes Schöning, Brent Hecht, and Werner Kuhn.
2014. Informing Online and Mobile Map Design With
the Collective Wisdom of Cartographers. Proceedings
of the Conference on Designing Interactive Systems
(DIS 2014), ACM Press, 765–774.
http://doi.org/10.1145/2598510.2598543
28. Beat Signer, Moira C. Norrie, Michael Grossniklaus,
Rudi Belotti, Corsin Decurtins and Nadir Weibel.
2006. Paper-based Mobile Access to Databases. Demo
Proceedings of the International Conference on
Management of Data (SIGMOD 2006), ACM Press,
763–765. https://doi.org/10.1145/1142473.1142581
29. Terry A Slocum, Robert B. McMaster, Fritz C. Kessler,
and Hugh H. Howard. 2009. Thematic Cartography
and Geovisualization. Prentice Hall, Upper Saddle
River, USA.
30. Bart Thomee, Ioannis Arapakis, and David A.
Shamma. 2016. Finding Social Points of Interest from
Georeferenced and Oriented Online Photographs.
ACM Transactions on Multimedia Computing,
Communications, and Applications 12, 2: 1–23.
http://doi.org/10.1145/2854004
31. Ying Zhang, Roger Zimmermann, Luming Zhang, and
David A Shamma. 2014. Points of Interest Detection
from Multiple Sensor-Rich Videos in Geo-Space.
Proceedings of the ACM Conference on Multimedia
(MM 2014), ACM Press, 861–864.
http://doi.org/10.1145/2647868.2655027
