54
R. Saameli et al.
map-making methods (Haworth and Bruce 2015). However, the full scope of opportunities offered by VGI is still underused by traditional humanitarian actors. In order
to understand this situation, Richards and Veenendaal (2014) have analyzed comprehensively the gap between the United Nations World Food Programme Crisis
Mapping Operations and Crowdsourcing Technology. They conclude that crowdsourcing captured a large amount of data, but not sufficiently the required ones for
the agency operations and with not the needed quality. Haworth and Bruce (2015)
also highlight this need of enhancing data quality assurance to enhance the relevance
of VGI for disaster management. Data quality in VGI relies heavily on Linus’s law,
which implies the more observers, the more likely an error will be identified. As
shown by Haklay et al. (2010) in their analysis of Open Street Map data quality,
this law seems to work well for spatial accuracy. However, Mooney and Corcoran
(2012) have also discovered serious quality issues with tags or annotate objects in
OSM. Such attribute data is usually much more needed than accurate geographical
coordinates to support humanitarian management.
Geo-information on health facilities in disaster areas are a good example of the
challenges of the use of VGI for humanitarian action. The most comprehensive
healthsite geodatabase based on VGI is probably OSM, but information on services
offered are still largely incomplete and questionable in terms of reliability. Other
health geodatabases with comprehensive set of helpful attributes for health workers
exist, but these databases are not easily shared outside of the health organizations
which have gathered them, or are only regional in their coverage. OSM and these
restricted datasets complement each other in terms of geographical coverage and
in terms of the information they contain, however they are almost never readily
available in a consolidated, freely and accessible way. Data exchange between VGI
communities and health organizations is usually unidirectional and punctual. Traditional humanitarian agencies tend to task digital communities only for specific tasks
lasting a rather short period of time (Burns 2014).
In order to address this issue, the Global Healthsites Mapping Project has been
launched in 2015 to create an online interactive map, Healthsites.io, of every health
facility in the world and make the details of each location and services easily accessible. A team of freelance developers, researchers, the International Committee of
the Red-Cross (ICRC), and the International Hospital Federation (IHF) have joined
their competences and networks in order to provide a single point of reference for
healthcare workers, aid agencies, contingency planners, government agencies, and
citizens who need access to a highly curated global dataset of healthcare facilities.
In order to meet this aim, the project team has to address three major challenges:
1. Integrate multiple unstructured datasets in one unique database
2. Enhancing the reliability of data
3. Foster sharing and updating of information
In Sect. 5.2, this paper will present the approaches that are currently developed
to address these three challenges. In Sect. 5.3, the paper will analyze the potential
impacts, risks, and the perspectives of this project.
R. Saameli et al.
map-making methods (Haworth and Bruce 2015). However, the full scope of opportunities offered by VGI is still underused by traditional humanitarian actors. In order
to understand this situation, Richards and Veenendaal (2014) have analyzed comprehensively the gap between the United Nations World Food Programme Crisis
Mapping Operations and Crowdsourcing Technology. They conclude that crowdsourcing captured a large amount of data, but not sufficiently the required ones for
the agency operations and with not the needed quality. Haworth and Bruce (2015)
also highlight this need of enhancing data quality assurance to enhance the relevance
of VGI for disaster management. Data quality in VGI relies heavily on Linus’s law,
which implies the more observers, the more likely an error will be identified. As
shown by Haklay et al. (2010) in their analysis of Open Street Map data quality,
this law seems to work well for spatial accuracy. However, Mooney and Corcoran
(2012) have also discovered serious quality issues with tags or annotate objects in
OSM. Such attribute data is usually much more needed than accurate geographical
coordinates to support humanitarian management.
Geo-information on health facilities in disaster areas are a good example of the
challenges of the use of VGI for humanitarian action. The most comprehensive
healthsite geodatabase based on VGI is probably OSM, but information on services
offered are still largely incomplete and questionable in terms of reliability. Other
health geodatabases with comprehensive set of helpful attributes for health workers
exist, but these databases are not easily shared outside of the health organizations
which have gathered them, or are only regional in their coverage. OSM and these
restricted datasets complement each other in terms of geographical coverage and
in terms of the information they contain, however they are almost never readily
available in a consolidated, freely and accessible way. Data exchange between VGI
communities and health organizations is usually unidirectional and punctual. Traditional humanitarian agencies tend to task digital communities only for specific tasks
lasting a rather short period of time (Burns 2014).
In order to address this issue, the Global Healthsites Mapping Project has been
launched in 2015 to create an online interactive map, Healthsites.io, of every health
facility in the world and make the details of each location and services easily accessible. A team of freelance developers, researchers, the International Committee of
the Red-Cross (ICRC), and the International Hospital Federation (IHF) have joined
their competences and networks in order to provide a single point of reference for
healthcare workers, aid agencies, contingency planners, government agencies, and
citizens who need access to a highly curated global dataset of healthcare facilities.
In order to meet this aim, the project team has to address three major challenges:
1. Integrate multiple unstructured datasets in one unique database
2. Enhancing the reliability of data
3. Foster sharing and updating of information
In Sect. 5.2, this paper will present the approaches that are currently developed
to address these three challenges. In Sect. 5.3, the paper will analyze the potential
impacts, risks, and the perspectives of this project.