1 colonne2 colonne3 colonne
1 ligne9.13.24.54
2 ligne2.48.89.65
3 ligne3.11.53.7
4 ligne4.39.026.2
Digital Revolutions: New Information Technology
Tools in 21st Century Politics
How OpenStreetMap respond to Disaster Crisis
Pierre Bland, Humanitarian OpenStreetMap Team Volunteer
Norvegian Center for Humanitarian Studies (CMI), Bergen, Norway,
2015-11-02
OpenStreetMap and HOT in Disaster Responses
Recent ActivationsContributorsObjects edited
Haiyan Typhoon, Philippines, Nov. - Dec. 2013
1,6004.5 million
West Africa Ebola outbreak, Mar.2014 - Mar.2015
3,31215.4 million
Vanuatu Archipel Typhoon, Mar.2015
3400.7 million
Nepal Earthquake, Apr-June 2015
7,12517.4 million
We support humanitarian teams in the disaster zones
for various Map needs: Humanitarians that deploy to the zone and
provide relief to the population
UN Agencies, International organizations, Logisticians that plan intervention, evaluate number of people affected, ressources to deploy
GIS specialists that support the organizations in the field
Humanitarian aid and Health teams travelling that need rapid access to various areas
Transport truck fleets, etc.
OSMA multitude of tools / products including on mobile devices
Flexible Organization / Coordination with DHN, Imagery providers, UN Agencies, in ternational organizations
Various expertise from developpers and OSM / HOT volunteers Requests We interface with ressources
Capacity to Crowdsource, to adapt to various situations and deliver rapidly Maps & Services of quality
Great opportunity in Philippines and Nepal disasters to coordinate with OSM local community
OSM Ecosystem, Coordination with PartnersOSM is a Remote effort essentially.
The DHN partners, other organizations and field teams to provide the Information gaps
Workflows to build together Data Collection, integration of OSM basemap in products
Trust, interconnections and orientation discussions to be developped both with partners and the Core Volunteers.
OpenStreetMap / HOT Services available immediatelyMaps & Services (Free and OpenData)
Data Exports for GIS analysis, Mobile devices Offline maps and road navigation (Android, IOS, GPS) Daily updates, more frequent when needed
HOT Softwares to coordinate crowdsourcing, export Data
Online Map and Road navigation, Paper maps, FieldPapers
Quality Control tools, Query Tools
Local communities to support
field teams -> Philippines and
Nepal experience
OpenStreetMap and HOT in Disaster Responses
Instant availibity of Maps & Services (Free and OpenData)
Data Exports Daily updates more frequent in the context of crisis and rapid updates of the database
OSM is a large Community of developpers and contributors
This is the civil society participating to support international organizations.
Adaptations from the organizattions were necessary on both side to adapt both to the Crowsourcing, coordination of volunteers through internet, outreach.learn how to work with others.
Adaptations were also necessary from the HOT Teams coordinating with UN agencies and international organizations with stricter rules of communication.
Quick Response to support humanitarian teams in the disaster zones Humanitarian OpenStreetMap Team Collective effort with the partners. Various technical and organizational aspects to take into account
Remote Volunteers coordinate with the Digital Humanitarian partners, Imagery providers, Damage assessment organizations, UN Agencies, International organizations, GIS specialists
Imagery acquisition and processing
Adapt the response to the context of each crisis
Crowdsource Remote Mapping / Monitor / Evaluate / Correct
Feedback, correctionsGIS specialists
People in the field
OpenStreetMap and HOT in Disaster Responses
A Mix of System + Maps & services readily available Expertise,Voluntarism, economic actors,Crowdsourcing effort
Long term viable? Important to manage tensions, to respect the Volunteers expertise, the partners contributions
Haiyan Response, Philippines, 2013Humanitarian organizations saw the potential of OpenStreetMap Maps & Services in Crisis Context. OSM / HOT adapted rapidly to provide various Poster maps, plus revising Map styles to show damaged buildings
OpenStreetMap and HOT in Disaster Responses
http://umap.openstreetmap.fr/en/map/hot-yolanda-haiyan-typhoon-activation_3628#8/11.558/124.887Red : Post-disaster, blue : pre-disaster
HOT / OSM community Activation for the Haiyan Typhoon, Nov 8, 2013
This map shows the crowdsource effort. The polygons correspond to the various Task Manager jobs to coordinate the remote mapping effort through internet.
OpenStreetMapHumanitarian Style
Transport Style ---> Minimal
West Africa Ebola outbreak, 2014, Zones covered by the OSM Response
uMap
http://umap.openstreetmap.fr/en/map/west-africa-ebola-outbreak-zones-covered-by-the-op_13842#6/8.081/-12.019This
map shows the crowdsource effort. The polygons correspond to the
various Task Manager jobs Coordination / Validation of the
Crowdsoruce remote mapping effort through internet.Phase 1: MSF
Switzerland + CartONG 3 towns context of fragile health
systems
- extend 2 border countries Sierra Leone + Liberia
- Imagery missing Coordination with Imagery providers- June,
restart after a pause Large urban areas to cover Monrovia,
Freetown, Conakry- Mid august UNMEER extension to cover regions in
the 3 countries most affected
West Africa Ebola Outbreak, March December 2014
MSF Switzerland - Mar 2014GIS specialists deployed with epidemiologists in the field
Acquisition of Imagery for the first 3 towns
Activate HOT to support contact case tracingQuick mapping of territories as the outbreak spread higways, villages, houses
Village names to facilitate locating people at risk
Free aerial imageryprovided by HIU US State Department, Airbus Defence and Space, MapBox helped complete Bing Imagery in areas not yet covered
Thousand of MSF / Red Cross volunteers travelling on the territory
UNMEER August 2014Deployment of international teams to support the development of the health sector and economic development in the 3 countries most affected
OpenStreetMap extends the mapping to large zones in these countries
OpenStreetMap and HOT in Disaster Responses
Medical staff, logisticians and thousand of of responders in the field that locate people at risk rely on Maps. Information and communication is part of the response.
One of the biggest challenge is tracing all contacts of an infected person > Tracing outbreaks with mapping and geolocation
Village names are necessary to locate rapidly the contact persons
OpenstreetMap Response - 2015 Nepal Earthquake
Google CrisisMap to monitor imagery
Previous ActivationsImagery, Post-Disaster Assessment, Coordination, experiments with drone Workflows / coordination to develop
Nepal Post-Disaster Imagery acquisition
Great effort of satellite company that re-oriented their satellite and took pictures day after day at this monsoon season.
OpenstreetMap Response - 2015 Nepal Earthquake
KLL traced Kathmandu years prior to the earthquake
04-25 Pre-Disaster Trace roads
04-28 Post-Disaster Kathmandu + Banepa IDP Camps
04-28 Pre-Disaster Trace Buildings Gorkha, Sindhupalchowk, Dunche, Langtang, etc
05-01 Post-Disaster outside Kathmandu Gorkha
05-01 Import place names
Timeline
Timeline
Pre-DisasterKathmandu valley was already traced by Kathmandu Living Labs
Day 1: Highway Network, mountain paths
Day 4: villages outline, buildings, Helico landing
Day 7: Place names
Post-disaster
Day 4: IDP Camps Kathmandu valley (2 days)
Day 11: IDP camps, damaged villages outside Kathmandu (jobs progressed slowly and are not completed)
OpenstreetMap Response - 2015 Nepal Earthquake
Crowdsourcing effort
Statistics of OSM Contributions
For the first week, an average of 1 million objects edited / day
and 1,000 contributors daily
In comparison, 1 million objects for the first month in Haiti, jan.
2010.
OpenStreetMap is an open system and anybody can open an account to
edit or organize a mapathon. The quality control / validation
process is essential to assure quality, accuracy of data.
OpenstreetMap Response - 2015 Nepal Earthquake
Long tail of contribution
The first 500 contributors contributed to 75% of the edits
4,850 contributors
(68% of contributors) participated only one day. They contributed
to 10% of objects edited.
Often, this was their first day of contribution.
OpenstreetMap Response - 2015 Nepal Earthquake
Features added to OSM
Profile, 04-24 06-07
Total correspond to 06-07
Objects added 04-25 to 06-07
Objects 04-24
OSM Edits in the six weeks of Activation
Building geometry often corrected New contributors trace imprecise
geometry
Indetermed Features represent only one percent of Data, but a high
propotion of objects deleted
Source: OsmPlanetStats, Pierre Bland
OpenstreetMap Response - 2015 Nepal Earthquake
Source: OSMPlanetStats, Pierre Bland
Prior to the earthquake, the Nepal OSM community added 48,000 km of roads. From Apr.25, the international OSM community and various support groups contributed adding an other 60,000 km to trace unclassified roads, tracks and paths in the remote areas.
Quality + Rapidity are essential to a good response
Crowdsource Mapping Earlier we correct problems, better it isTutorials
Clear instructions Jobs for beginners
OSM Editors adapted to new contributorsFunctions that facilitate edit specific features (ie Building geometry + Tags), Side Instruction Panel, etc.
Validation function inside the Editor that the contributor understand easily
Tasking Manager validation limited to experienced mappers
Assure a better workflow controlMonitoring contributors, identify early the problems
Other Validations including Semantic
OpenStreetMap and HOT in Disaster Responses
OSM Semantic Ontologic precisionAttributes are not fixed. This let place for constant innovation. But represent also a challenge to assure that features are identified by the various applications
OpenStreetMap and HOT in Disaster Responses
Ontologic precision vs OSM Feature Key Value pairs
Nearly 1,000 Common Key-Value pairs represent OSM features
(Map_features wiki)99% of the objects in OSM Nepal 2015-04-24 and
2015-06-07 can be represented as Features since they use a Common
Key-Value
(The Visible part to softwares)
OpenStreetMap and HOT in Disaster Responses
1% -The Invisible part Syntax error, no attributes, errors, or simply notes added Examples Highway=residential (Uppercase)
highways=residential (s suffix)
highway=Blvd xxx (a name in the Value Field)
Damage=I think that this building has collapsed,
Etc.
Semantic analysis of the OSM Planet data - 04-24 and 06-07
Various studies confirmed the good quality of the OSM dataGood results even in emergency context
To respond quickly and efficiently, the earlier we can correct such data, the better it is Tutorials, clear instructions, Editors adapted to new contributors.
Various Validation, Qualiy control tools
Semantic analysis to complete these controls
Semantic Analysis of OSM Tags Bring the Invisible into Light
JOSM Editor > MapStyle to Test integration of semantic
validation To switch the focus from the Visible to the Invivisble
portion of Data NoFeature Validation Style highlights Objects with
no Valid Feature
(ie (highway, building, landuse, office, shop, etc)
OpenStreetMap and HOT in Disaster Responses
NoFeature Validation MapStyle
Appendix Various examples
West Africa Ebola outbreak Detailed mapping of the urban
areas
Before / After map of Monrovia
http://pierzen.dev.openstreetmap.org/hot/leaflet/OSM-Compare-before-after.html#14/6.3334/-10.7868
OpenStreetMap and HOT in Disaster Responses
OpenStreetMap and HOT in Disaster Responses
Nepal Earthquake, 2015
Map from USGS shows epicentre and aftershocks
on an OpenStreetMap (Mapquest) Relief Basemap
OSM integrated in various products
Gis Data Marts Layers of various features
OSM query tools integrated in GIS softwares
http://yolandadata.org/maps/new?layer=geonode:damage_lines
Yolanda Content Management From Geonode platform
Explore, Export Maps
IOM personal Joe Lowry CCBYSA2.0 http://flic.kr/p/hHMxee
Poster Maps, City index
American Red Cross. Used with permission https://twitter.com/RedCross/status/401088520481042432
OsmAnd Offline Navigation
on small devices
Presentation available from slideshare.net/pierzen
% of create% of delete% of modify
Building0.5373832429101660.3450098997760380.709778956289976
Network0.2877458982699650.1981758445464680.146081386380917
Nature0.0694449782545880.03739194096920430.0210841116146737
Landuse0.07028065040843250.1227187537943130.0240154411744667
Damage0.01575492748247160.0147685183515520.0163752483369261
POI's0.002497719014004960.02544488884240540.00600437127494611
Other0.0001177188631959660.001426962253422240.000557650957510881
Ind0.01677486479717690.2550631914665970.0761028339705838
04-24'04-24 06-07
Building11209497071205
Networks17028013771483
Nature1156585901993
Landuse148258810658
Damage313225918
POI's623399585
Other4044313437
Ind2063326489
Process and Host Imagery (TMS Server). Used in OSM Editors as background image to trace the buildings and roadsHumanitarian mapping workflow : Revise the Tagging schema for damaged buildings and infrastructuresExtracts : Provide frequent updatesTools and Services : Revision to show Damaged buildings Online nd Paper maps.Wednesday 13th Nov : Starts Post-disaster Mapping (damaged buildings and road blocks) of Tacloban. Friday Nov. 15th : In week, more then 900 OpenStreetMap mappers have contributed to this response. They have modified more than 2 million objects on the map (1.3 million for Haiti)
C. Heipke, Crowdsourcing geospatial data, ISPRS Journal of Photogrammetry and Remote Sensing, vol. 65, pp. 550-557, Nov 2010.