The Urbangene Project

Experience from a Crowdsourced Mapping Campaign

Jens Ingensand

, Marion Nappez

, St

ephane Joost

, Ivo Widmer

, Olivier Ertz

and Daniel Rappo

GIS Lab, Department EC+G, University of Applied Sciences Western Switzerland, Yverdon-les-Bains, Switzerland

Laboratory of Geographical Information Systems (LASIG), School of Architecture, Civil and Environmental Engineering

(ENAC), Ecole Polytechnique F

erale de Lausanne, Lausanne, Switzerland

Department Comem+, University of Applied Sciences Western Switzerland, Yverdon-les-Bains, Switzerland

{jens.ingensand, marion.nappez, olivier.ertz, daniel.rappo}@heig-vd.ch, {stephane.joost, ivo.widmer}@epﬂ.ch

Keywords:

Citizen Science, Crowdsourcing, Participatory WebGIS, Social Media, Urban Biodiversity, VGI.

Abstract:

Geospatial crowdsourcing applications are emerging systems that enable researchers to collect important in-

formation that otherwise would be difﬁcult to obtain. In biodiversity monitoring, crowdsourcing is a promising

approach as it beneﬁts from a large group of people with an often underestimated biodiversity and taxonomy

knowledge. Despite its huge potential, crowdsourcing approaches are still underrepresented in biodiversity

monitoring. We here evaluate a participatory crowdsourcing web mapping platform that was developed to get

information about geographic locations and biodiversity characteristics of urban ponds in the Geneva cross-

border region. An important fraction of urban ponds is assumed to be located on private grounds, which

makes the participatory crowdsourcing approach very valuable. A media campaign was initiated, encouraging

citizens participate and to digitize ponds. In this paper we a) evaluate and discuss the impact of the media

campaign on the usage behaviour and history of citizens using the crowdsourcing platform and b) assess the

quality of the digitized data that has been collected. This study shows that through media campaigns, citizens

can be mobilized and motivated to participate in biodiversity crowdsourcing projects. Results indicate that

large quantities of users were recruited through social media. However, only a small fraction of about 3% of

the mobilized people digitized ponds on the platform. The majority of these users (68%) digitized one pond

while 32% digitized two or more ponds. This study shows that it is important for crowdsourcing platforms

to be designed and planned in order to facilitate its usage. However, it is crucial for the success of such cam-

paigns to offer something in return to the users and to encourage them to interact among themselves. We

suggest that future crowdsourcing biodiversity mapping campaigns should have mobile-optimized interfaces.

Mobile devices have the potential to e. g. automatically register coordinates for biodiversity sighting and for

uploading respective pictures directly in the ﬁeld.

1 INTRODUCTION

Spatial crowdsourcing systems are an emerging type

of systems. These systems have their roots in the

experience from PPGIS (public participatory GIS,

Sieber (2006)) and many non-spatial crowdsourcing

projects, such as wikipedia (wikipedia.org; Wikime-

dia Foundation Inc., San Francisco, USA). Accord-

ing to Goodchild and Sui (2011), there are two con-

verging trends that also contribute to the success of

spatial crowdsourcing systems: geographic informa-

tion systems are increasingly social and online social

networking sites such as Facebook are increasingly

location-based. The success of the OpenStreetMap

project (openstreetmap.org; Open Data Commons

Open Database License) is an example of a spatial

crowdsourcing system where data can potentially be

collected and used by everybody. Bartoschek and

Kessler (2013) state that most publications related to

volunteered geographic information (VGI) focus on

this project.

A key challenge is to get people to use a crowd-

sourcing system or application and to participate in

a mapping campaign. This challenge can be divided

into two components - on the one hand the targeted

citizens need to ﬁnd the application and on the other

hand, citizens must want to use the system once dis-

covered. The incitement to participate in crowd-

sourced mapping campaigns is often given by existing

communities such as, for instance, the global ornitho-

logical network eBird (Robbins, 2013) or the crisis

mapping tool Ushahidi which has been successful for

178

Ingensand J., Nappez M., Joost S., Widmer I., Ertz O. and Rappo D..

The Urbangene Project - Experience from a Crowdsourced Mapping Campaign.

DOI: 10.5220/0005468501780184

In Proceedings of the 1st International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM-2015), pages

178-184

ISBN: 978-989-758-099-4

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

mapping observations of violence (Liu et al., 2010).

Both projects, Ushahidi and eBird, have one thing

in common - the participants motivation is already

given through the fact that there is a common inter-

est to collect data and to visualize and analyze data

that has been collected by others.

However, if there is no existing community avail-

able, scientists who want to collect speciﬁc data

through a crowdsourced mapping campaign face a

challenge in getting people to use their system. First

of all, potential participants need to be informed of

the system’s existence e.g. through media campaigns

or through social networks. Secondly, the system that

is used needs to be optimized for the tasks of sharing

and editing spatial data. According to Haklay (2013),

it is important that the equipment that volunteers uti-

lize in order to gather information is adapted to their

skills and personal knowledge.

Aquatic ecosystems in urban areas are inﬂuenced

and threatened by human activity, which often causes

fragmentation, degradation or loss of environments

and reduces connectivity between remaining habitat

patches. Ponds and streams are important biodiver-

sity hotspots in urban environments. In order to allow

a spatially explicit conservation planning for urban

aquatic ecosystems, with the goal to maintain, im-

prove and connect urban aquatic habitats, it is thus

important, in a ﬁrst step, to locate and characterize

these vulnerable environments. We evaluated a par-

ticipatory web mapping platform, which allows for

the input and retrieval of information about the loca-

tion and characteristics of urban ponds. An impor-

tant fraction of ponds is assumed to be located on pri-

vate grounds, which makes the participatory crowd-

sourcing approach very valuable, as such information

would be difﬁcult to obtain otherwise.

The platform, called Urbangene, allows citizens

to digitize ponds and to report speciﬁc species that

have been identiﬁed in or around these ponds. It was

developed in the framework of the biodiversity re-

search project Urbangene, giving its name to the plat-

form. The main goal of the Urbangene project is to

investigate the impact of urbanization on biodiversity

an thereby to allow for more sustainable urban devel-

opment and the conservation, management and pro-

motion of biodiversity in urban areas.

Urban biodiversity has both a direct and an in-

direct value for a city’s residents; it is for instance

important for the physical and psychological health

(Aldous, 2007; Fuller and Irvine, 2010). However,

urban aquatic biodiversity (such as ponds and its in-

habiting species) is less apparent to the citizens than,

for example, bird biodiversity. Moreover the map-

ping of observations of violence Liu et al. (2010)

seems to have a more direct gain to the residents of a

city. Thus, developers of the Urbangene platform ﬁrst

had to call the attention of the residents to the urban

aquatic biodiversity problem through a media cam-

paign. Developers optimized the platforms interface

following the most important interface design guide-

lines (e.g. Shneiderman’s eight golden rules of in-

terface design (Shneiderman and Plaisant, 2009)) and

developed several mock-ups in order to test and dis-

cuss multiple design options. A challenge in estab-

lishing the design of the system was to identify the

potential end user of the platform.

In the following sections we will focus on the Ur-

bangene biodiversity mapping platform and the media

campaign that has been started in order to encourage

volunteers to digitize ponds. Thereafter we will ana-

lyze the results of the mapping campaign and respond

to the research questions:

1. Which medium was more successful in encour-

aging citizens to use the system? The type of

medium can give important indications for further

campaigns.

2. Is it possible to characterize the users of the plat-

form? When were they using the system? Where

were they located? How many times were they ac-

cessing the system? A user proﬁle can help further

crowdsourcing campaigns to improve user inter-

faces and to adjust the systems to the users needs.

3. What is the quality of the digitized data? How

accurate are the results? Is there any evidence of

fraud? This question is important since it tells if

users were willing to collect usable data and if the

tool was adapted to the users skills and expecta-

tions.

Finally we will discuss the results and give our inten-

tions for further work.

2 THE URBANGENE PROJECT

AND PLATFORM

The main aim of the Urbangene project is to investi-

gate the impact of urbanization on biodiversity in the

Geneva cross-border area with the goal to allow for

more sustainable urban development and the conser-

vation, management and promotion of biodiversity in

urban areas. The Geneva cross-border area (Greater

Geneva area) includes urban, suburban and rural terri-

tories in both Switzerland and France around the city

of Geneva.

The main objectives of the Urbangene project are

thus the assessment of the current state and the dy-

namics of biodiversity, the quantiﬁcation of landscape

TheUrbangeneProject-ExperiencefromaCrowdsourcedMappingCampaign

179

elements on gene dispersal and functional connectiv-

ity of both animal and plant species and, further, the

identiﬁcation of factors that inﬂuence gene ﬂow and

adaptation of species in urban areas.

In order to collect data, the web-based Urbangene

platform had been developed. The focus of this data

collection platform was on the identiﬁcation and lo-

calization of species such as toads and newts, and also

on their habitats: ponds. Due to the fact that ponds of-

ten exist on private property, an important point was to

conceive the platform as a crowdsourced data collec-

tion system, and thereby to implicate the population

of the area. The main function requirement was the

digitization of ponds and the speciﬁcation of species

that inhabit the ponds. Users would be able to see the

ponds and species digitized by other users and would

thus be encouraged to digitize other ponds.

On the client side, the Urbangene platform was

developed using HTML5/CSS3 technologies. Map-

ping functionality was implemented through the

JavaScript library Leaﬂet and enhanced by the jQuery

JavaScript library and its numerous extensions. On

the server side Apache, PHP and the database sys-

tem PostgreSQL/PostGIS were used for receiving and

storing user input.

Much emphasis has been put on the design of the

user interface in order to facilitate the task of digitiz-

ing ponds; e.g. an adaptive questionnaire has been

implemented that shows questions depending on the

answers that the user gave to previous questions or

messages that appear in order to guide the users. Fig-

ure 1 shows a screenshot of the platform.

Figure 1: A screenshot of the Urbangene platform.

Once the user gets to the web-based Urbangene

platform, a welcome screen is shown. The user then

proceeds to the map where the already digitized ponds

and the limits of the Greater Geneva area are shown.

The system then indicates through a message that a

new pond can be added through sliding a pond-icon

on the map. Thereby the location of a pond is digi-

tized as a point feature (Figure 1). Once the user has

put the icon on the map, the system requires the user

to zoom and to check the pond’s position. At the same

time, the questionnaire is shown and the user can ﬁll

in the questions, such as:

• the species identiﬁed in the pond (sample images

were shown, see Figure 1)

• the distance to a natural space from the user’s res-

idence

• how much time a week the user spends in nature

Another important question was whether the user

would allow scientists to tissue samples in order to

genotype the animals in the case where the pond was

located on their property. In addition, the platform al-

lows the user to upload pictures of observed species

which could not be clearly identiﬁed, and need to be

cross-checked by biodiversity experts. Once the ques-

tionnaire is completed, the user’s feedback is stored in

the database and its pond icon is made visible on the

map.

3 MEDIA CAMPAIGN

In order to call the citizens attention to the urban bio-

diversity thematic and to encourage them to partici-

pate in the pond survey, a media campaign was started

by the university’s media and communication center

and the creators of the Urbangene project. The me-

dia and communication center sent a press release

to a large number of newspapers and radio stations

throughout Switzerland. A project page and a Face-

book page were created, and the link to the platform

was published on web pages and forums.

The ﬁrst articles about the Urbangene platform ap-

peared in four newspapers on the 20th of March 2014;

four articles appeared the day after and the national

Swiss Radio dedicated a short radio transmission to

the project. Overall, news about the Urbangene plat-

form appeared 13 times on Swiss media (see Figure

2).

4 DATA COLLECTION

In order to answer our research questions, we did

not only use the platform’s database that contained

the digitized ponds and the responses to the ques-

tionnaire, but also the system’s web server log. The

Apache log ﬁle contains useful information such as

the users’ ip-addresses, the time and date users ac-

cessed the system, the user’s browser type and the

URL from which the user arrived at the system.

GISTAM2015-1stInternationalConferenceonGeographicalInformationSystemsTheory,ApplicationsandManagement

180

Figure 2: Extract from a newspaper article encouraging

readers to digitize ponds on the Urbangene biodiversity

mapping platform. (Tribune de Geneve, April 2, 2014).

The user’s ip-address can give information about

the Internet provider of the user and thus if users were

accessing the system using a wired Internet connec-

tion or through a mobile 3G or 4G network. More-

over ip2location databases (IP2Location.com, Pulau

Pinang, Malaysia) can give an approximate location

(e.g. at zip-code level) of the ip-address. The user’s

browser type allows to analyze if the user was using

a mobile device such as a tablet computer or a smart-

phone; in some cases it also gives information about

the operating system installed on the device (e.g. type

and version). The web page the user was visiting

before clicking on the platforms URL (the referrer)

gives an indication how the user found the system;

e.g. through a search-engine, through a newspaper’s

web page or through social media.

We parsed the Apache log ﬁle into a database and

aligned its data to the data that had been stored in the

database through the platform. This new database was

this study’s main source of information.

5 RESULTS

5.1 Data Collected

The data collected on the Urbangene platform

stretches over a period of ﬁve months; from the 20th

of March until end of August 2014. The main data for

this analysis can be characterized as follows:

• After cleaning the log ﬁle for robots, test users and

other irregular data, we were able to identify 933

distinct IP addresses that used the system during

the given period of time.

• 53 ponds were digitized within the limits of the

Greater Geneva area.

• 34 distinct users digitized the 53 ponds.

• Most digitized ponds were located in and around

the city of Geneva (within the borders of Switzer-

land). Only six ponds were digitized in France.

Figure 3 shows the locations of the ponds.

Figure 3: Map showing the results. Blue dots: ponds that

have been located and described by citizens using the Ur-

bangene biodiversity mapping platform. Pink area: the

French part of the Greater Geneva area, Orange area: the

Swiss part of the Greater Geneva area.

5.2 Effectiveness of the Media

Campaign

In order to measure the effectiveness of the media

campaign, we analyzed the referrers in the log ﬁle

(i.e. the page the user was visiting before connect-

ing to the platform). Figure 4 shows the results of

this analysis. Twenty-ﬁve percent of the users reached

the platform through the project’s homepage, 12%

through a media’s (newspaper or radio station home-

page) and 11% through a search engine. These num-

bers added together suggest that almost half of the

users (48%) found the system through either news-

paper or radio. 18% of the users found the URL of

the system through other homepages (e.g forums, etc)

and 34% of the users came from social networks such

as Facebook and Twitter. The high number of social

media users was surprising since the Facebook page

TheUrbangeneProject-ExperiencefromaCrowdsourcedMappingCampaign

181

only had about 30 followers at the time of writing this

paragraph.

The fact that only six ponds were digitized in

France could be an indication that the media cam-

paign did not sufﬁciently reach many citizens living

in France - however, the part of the Greater Geneva

region that is situated in France is partly mountain-

ous and much less populated than Geneva, the second

most populated city in Switzerland.

Figure 4: How the users found the platform, i.e. last Inter-

net pages users visited before connecting to the Urbangene

biodiversity mapping platform.

We also analyzed on which days the users used

the system and when the users digitized ponds. Fig-

ure 5 shows clear peaks at around the 20th and 21st

of March. Interestingly most of the ponds were digi-

tized a few days after the ﬁrst articles appeared. These

results lead to the argumentation that the media cam-

paign quickly motivated many users to take a look at

the system; however, the interest faded after the ﬁrst

week. The observation that users still digitized ponds

some days after the ﬁrst peaks appeared, might be in-

terpreted that these users either were looking at the

newly indicated pond locations for the occurrence of

different species that were demanded in the question-

naire or went home after discovering the platform and

then digitized a pond.

Figure 5: Number of different IP-addresses that accessed

the Urbangene biodiversity mapping system and number of

digitized ponds from March to August 2014.

5.3 Characterization of the Users of the

System

Most users (59%) who accessed the system (we are

referring to the 933 distinct IP addresses) were using

a desktop or laptop computer, 8% a mobile system

such as a smartphone or a tablet computer and 33 %

were using a non-identiﬁed system.

Figure 6: Number of platform requests according to the

time of day.

Most users used the platform during ofﬁce hours

with peaks at 10 a.m and 2 p.m. (see Figure 6); an

analysis of the Internet provider of the users showed

many different companies. A possible interpretation

of these numbers is that people used the system at

their workplace, after drinking a coffee and reading

the newspaper at 9 or 10 a.m. or after having lunch.

This hypotheses can be strengthened by an analysis of

the day of week people accessed the system - the data

showed that on average only 9% of the trafﬁc gen-

erated on the site occurred on a Saturday or Sunday,

while 15% - 18% occurred on a working day. The

vast majority of users (68%) digitized one pond; few

users digitized two or three ponds and only one user

digitized six ponds. (see Figure 7).

The ip2location database that we used for iden-

tifying the position of the users gave an interesting

result for the users who digitized ponds: nine out of

34 users (26%) were located in the canton of Geneva,

three in France and seven in other Swiss cantons. 15

users (44%) were using a mobile broadband connec-

tion and it was thus not possible to identify a more

precise location than Switzerland. A presumably high

percentage of mobile device users was also conﬁrmed

when we analyzed the browser type of the users who

digitized ponds - at least seven users (21%) were us-

ing a tablet computer.

More than half of the users who digitized a pond

(18 out of 33) answered that a natural space can be

found around their house; 5 out of 33 replied that na-

ture is close to where they live (some hundred me-

GISTAM2015-1stInternationalConferenceonGeographicalInformationSystemsTheory,ApplicationsandManagement

182

Figure 7: Number of digitized ponds per user.

ters) and seven people did not answer the questions

at all. 10 people were spending at least one hour in a

natural space a day, while 13 were spending at least

some hours a week. The most common species that

had been identiﬁed was the common toad (bufo bufo)

(found in 33 ponds). The common frog (rana tempo-

raria) had been identiﬁed in 17 ponds and the alpine

newt (ichthyosaura alpestris) in 16 ponds.

5.4 Quality of the Digitized Ponds

The locations of the 53 digitized ponds were com-

pared to other data sources in order to verify its ac-

curacy. For the Swiss part of the Greater Geneva area

we used a pond database that had been established

by the Geneva Institute of Technology, Architecture

and Landscape (hepia). Moreover, we used Open-

StreetMap data and Google Maps imagery in order

to identify ponds that were not contained in the refer-

ence database from hepia.

These three reference data-sources conﬁrmed the

location of 26 ponds (49%), 19 locations were not

conﬁrmed and thus possibly errors (36%) and 8 lo-

cations (15%) were either not far from a conﬁrmed

pond (and thus perhaps only an imprecision), not vis-

ible due to presumably covering trees, or objects that

are other water objects (e.g. fountains or swimming

pools)

6 DISCUSSION

We here analyzed the impact of a media campaign on

the usage behavior and history of citizens using the

Urbangene mapping platform. Although the area of

investigation was located in Switzerland and France,

the media campaign focused mainly on Switzerland.

This fact possibly explains why only few ponds were

digitized in France. The impact of the media cam-

paign could be seen in the fact that several hundred

people visited the mapping platform.

However, in relation to the effort of the media

campaign, the fact that only 3% of the users digitized

53 ponds might seem poor, moreover the trafﬁc on the

site quickly declined after only a few days. Having

this in mind and along with the fact that most users

only digitized one single pond leads to the conclusion

that most users did not come back to check how the

project was evolving.

There are many motivational reasons for one user

to become a contributor (Shneiderman and Preece,

2009), yet one reason explanation for our observa-

tion might be found in Haklay (2013) who states that

public participation can be categorized into four lev-

els ranging from crowdsourcing (i.e. citizens that are

used as sensors) on the lowest level, distributed in-

telligence, participatory science and extreme citizen

science (e.g. citizen participating in the problem deﬁ-

nition) on the highest level. In crowdsourcing citizens

are thereby used as data collectors and the sum of the

digitized data are not of direct interest to the users.

On the other hand, many successful crowdsourc-

ing systems such as OpenStreetMap offer something

in return for the work of the users: this is the possibil-

ity for one individual to reuse the data added by many

users. Another key for successful crowdsourcing bio-

diversity campaigns appears to be related to the level

of networking and organization of communities that

are interested in speciﬁc species. These communities

are often already well-organized and established and

in such cases crowdsourcing tools help to gather data

that the whole community is interested in.

The fact that many users found the platform

through social media is encouraging. The effort to

arrange Facebook pages or to tweet URLs is less than

starting a campaign including print media and radio

stations. Furthermore it is less demanding for the user

to click on a URL that is communicated through so-

cial media (e.g. a Facebook page) than retyping an

address that is printed in an article or communicated

during a radio transmission.

7 CONCLUSIONS AND FURTHER

WORK

This study indicates that through media campaigns,

many citizens can be mobilized and motivated to par-

ticipate in biodiversity crowdsourcing projects.

Even though it is important for crowdsourcing

platforms to be designed and planned in order to fa-

cilitate its usage, what is more crucial for the suc-

cess of such campaigns is to offer something in return

to the users. This would encourage users to interact

more among themselves in checking the data added

TheUrbangeneProject-ExperiencefromaCrowdsourcedMappingCampaign

183

by other users and possibly in adding new data to the

platform in the future.

Another important lesson learned is that social

media is able to recruit large quantities of users with

relatively little effort.

We suggest that future crowdsourced mapping

campaigns should have mobile-optimized interfaces.

Mobile devices such as smartphones or tablets are

promising tools for biodiversity monitoring during

outdoor activities. Applications for such systems have

the potential to for example automatically register co-

ordinates for biodiversity sighting and for uploading

respective pictures directly in the ﬁeld. The usage

of mobile devices is increasing and users could be

recruited to spend some time using a crowdsourced

system during their journey to work or other activities

during which they could use their mobile devices. An-

other advantage is that most social media platforms

such as Facebook or Twitter are available on mobile

devices. Thereby users could be recruited through

mobile social media applications and then directed to

crowdsourced mapping platforms without the need to

switch device.

ACKNOWLEDGEMENTS

The authors would like to thank all the participants

who used the Urbangene platform and the Geneva

Institute of Technology, Architecture and Landscape

(hepia) for providing the pond-database.

REFERENCES

Aldous, D. (2007). Social, environmental, economic, and

health beneﬁts of green spaces. In Proceedings of the

International Symposium on Horticultural Plants in

Urban and Peri-Urban Life, pages 171–85.

Bartoschek and Kessler (2013). Crowdsourcing Geo-

graphic Knowledge, chapter VGI in Education - from

K-12 to Graduate Studies, pages 341–360. Springer.

Fuller, R. and Irvine, K. (2010). Interactions between peo-

ple and nature in urban environments. Urban Ecol-

ogy; Gaston, K .J., Ed., Cambridge University Press:

Cambridge, UK,, pages 134–171.

Goodchild and Sui (2011). The convergence of GIS and so-

cial media: challenges for GISscience. International

Journal of Geographical Information Science (IJGIS),

25(11):1737–1748.

Haklay, M. (2013). Citizen science and volunteered geo-

graphic information: overview and typology of par-

ticipation. In Sui, D., Elwood, S., and Goodchild,

M., editors, Crowdsourcing Geographic Knowledge:

Volunteered Geographic Information (VGI) in Theory

and Practice, pages 105–122. Springer.

Liu, S., Iacucci, A., and Meier, P. (2010). Ushahidi Haiti

and Chile: next generation crisis mapping. ACSM Bul-

letin, 246.

Robbins, J. (2013). Crowdsourcing, for the birds. The New

York Times.

Shneiderman, B. and Plaisant, C. (2009). Designing

the User Interface: Strategies for Effective Human-

Computer Interaction (4th Edition). Pearson Addison

Wesley.

Shneiderman, B. and Preece, J. (2009). The reader-to-leader

framework: Motivating technology-mediated social

participation. AIS Transactions on Human-Computer

Interaction, 1(1):13–32.

Sieber, R. (2006). Public participation geographic infor-

mation systems: A literature review and framework.

Annals of the Association of American Geographers,

96 (3):491–507.

GISTAM2015-1stInternationalConferenceonGeographicalInformationSystemsTheory,ApplicationsandManagement

184