Last night I responded to the Humanitarian OpenStreetMap team’s call to action to join the global effort to improve the state of maps in the Philippines, in the wake of the ongoing humanitarian crisis there. Up to date maps are needed by relief organisations such as the Red Cross and Doctors Without Borders to help them navigate the area and allocate resources.This mapping session was as revealing as it was informative for me, and there are a couple of points I’d like to share.

Harry Wood and machine learning researcher Dan Sowell taught me how to make my first edits to the OpenStreetMap (OSM) data using a combination of the JOSM editor (you can find tutorials here) and the OSM Tasking Manager created by the Humanitarian OSM team. I began by tracing out buildings on the satellite imagery (provided by DigitalGlobe). While much of the transport backbone has already been traced in recent days, tracing buildings helps improve population density estimates. Other ongoing mapping tasks include tracing out the newly acquired post-disaster imagery.

Ivan Gayton (Head of Mission at Medicins sans Frontiere/Doctors Without Borders) gave an impromptu talk in which he described the importance of disease mapping which relies on three sources of data:

  • Tracing work into OSM by volunteers worldwide.
  • Spreadsheets of disease diagnoses by doctors giving street-level or village-level records and times.
  • Surveying by local communities who can provide the information to connect the OSM data with street names or village names, to enable geocoding (the conversion of addresses to longitude and latitude on the map) and subsequently data analysis.

Using disease maps, the carriers of the disease can be inferred from spatial and temporal patterns. For instance, Ivan described how some animal-borne disease tends to spread out in a circular radius, while some human-borne disease may be found in more localised clusters, or tracking along transport networks. Knowing the distribution and activity level of the disease helps to optimise resource allocation, for instance, aiding decisions on whether vaccination or a measure such as water chlorination will be more effective.

This kind of disease mapping has its roots in the celebrated cholera map made by John Snow in 1854 around the Broad Street pump. My understanding is that continuing efforts to scale this method could improve the range of predictive or preventative measures that can be taken.

Thanks to Harry Wood for organising and the Open Data Institute for hosting the event. The mapping is still ongoing and you can get involved either by joining the tracing effort or donating to the Humanitarian OSM team to support their work. We look forward to finding ways to work with the Humanitarian OSM team in the future.