[talk-au] Interesting take on OSM growth and expansion...

John Smith deltafoxtrot256 at gmail.com
Wed Apr 28 15:42:01 BST 2010


“What would it take to map an entire country?”

With the growing visibility of Map Kibera, that question is coming
more frequently, especially in Africa, where both OpenStreetMap and
traditional mapping are widely absent. This is a massive question,
which is going to depend very much on circumstances of that country,
and on who is asking that question; and in the end may be better
answered by a different question. In response to a couple queries,
from Liberia and Malawi, I decided to write up a few blog posts to
start off those conversations, and serve as reference for any of the
other 200+ countries on this planet. To start, going look at a few
examples to serve as models for answering the question.

Up front, the question assumes one very important thing; the
historical growth pattern of OSM isn’t happening. Traditionally, a few
individuals had their minds blown by a conference presentation on OSM,
or maybe a random blog post somewhere, and they start mapping their
home town. And when that looks to be a big task, and they start eyeing
the next town, they start recruiting others through mapping parties. A
mailing list is set up. The virus starts to spread, and OSM might get
the attention of a local government or two, maybe some companies, and
soon, the country is well on its way to being mapped. The growth is
organic. It might take years. The preconditions are important.
Roughly, there has been an active community already of technically
proficient people who have leisure time, perhaps already contributing
to open source projects. In other countries, there may not be a
technical community, or socioeconomic conditions make leisure time a
valuable and scarce. Other places may be in conflict. These are places
where people start to consider intentional interventions to get
mapping going.

The first place I took part in an intentional effort was India. Now,
India is well known for having a broad and highly skilled technical
base, but at least in 2008, there was proportionally small open source
community. There had been interest in mapping through the Free Map
project, and Schuyler and I took the invitation to promote OSM through
a slightly insane schedule of mapping parties in 7 Indian cities in
one month. A wonderful experience. But it did not instantly translate
into a frenzy of mapping activity. The idea gestated for a time, and
slowly, individuals took up the cause, and now the OSM India community
is vibrant, centered in cities. One place to note is Mumbai, despite
extensive interest, is still not active, largely due to an import of
AND data that turned out to be much less than accurate, discouraging
further editing. An issue we’re still looking at, and generally, I’ll
talk about existing data sources and imports in another post.

Palestine was a very intentional mapping, first of the entire West
Bank, and then of Gaza. JumpStart International funded a project to
create a complete, public domain map of all roads. We estimated a time
of about 6 months, which turned out to be accurate (more on time
estimates in another post). The approach taken was to hire engineering
graduates, in teams based in each region, train them through mapping
parties, and coordinate the incoming data. Gaza was not to be mapped,
but the 2009 crisis there motivated the remote OSM community to fund
raise to purchase recent satellite imagery of the entire strip, and
trace remotely. After the conflict was over, JumpStart entered Gaza
and built up a team to enhance that work with a combination of GPS
tracing and imagery. The result is a complete road map of Palestine.
However, no sustained OSM community of interested individuals, local
companies, local government, and UN agencies is left. The individuals
involved saw this as temporary paid work, and many of them have taken
jobs outside Palestine. The wider community wasn’t engaged with an aim
to build capacity. So now, even though Israeli OSM has been interested
to hold joint activities with Palesinte, and Ramallah now has street
names, the map isn’t being updated.

In Kenya, we haven’t set out to map the entire country, but it seems
like the base we’ve built in Kibera is ready to spread throughout
Nairobi and beyond. We’ve focused heavily on Kibera, but with the idea
always that the group here, and the entire community, will be able to
take the project forward. The impact of mapping a place that was
unmapped, and considered unmappable, has made a great impression
within civil society, the government, and the UN. We’ve taken a lot of
time to outreach to others already working with map data. The approach
is more like planting a single seed, nurturting it, and then allowing
it to grow. It’s yet to be seen how this pans out.

JumpStart started next working in Georgia. Their approach is long term
but different. They have first focused on conflict areas, following
the Russian Georgia conflict in X, and also in the capital, Tbilisi.
>From this start, they are building a country wide network of regional
offices, Open Maps Caucasus, supported by an NGO structure. They are
producing curriculum, hiring a team, and doing extensive outreach with
map data users. The question to my mind is the long term
sustainability of the structure. If JumpStart decides to stop
supporting the NGO structure, will it be able to find more funding to
keep going? And if that doesn’t materialize, will a community be in
place to take it forward organically? Again, yet to be seen, and a
question they are certainly considering.

In the USA, mapping hadn’t taken off until last year. This was perhaps
due to the TIGER data set, available public domain. Atlanta mapping
weekend, involving hundreds of people in cooperation with the Atlanta
municipal government. However, after a time CloudMade discontinued the
ambassadors program. From there, what has developed has looked like a
more traditional growth, with a wide spectrum of individuals,
companies, organizations, and government getting interested and
involved. The GeoDC group has been very active, and recently OSMF US
has incorporated. Hard to calculate what effect the intentional effort
of CloudMade had on this growth, though it was certainly a

Finally, many have seen the extremely rapid growth of the map in
Haiti. Prior to the quake, very little data existed there for familiar
reasons. Following the disaster, and the release of imagery for
derived works and other data sources, remote mappers quickly and
spectacularly produced comprehensive maps of Port au Prince and much
of the country. Mappers wanted to contribute to the relief effort.
Certainly in other unmapped places, remote mappers are motivated to
contribute just out of interest. If you look at any capital city
covered by Yahoo imagery, there will have been remote mappers
contributing at least geometries. Does this translate to places not in
crisis, but in more prolonged issues that prevent mapping? Probably to
some extent, but not nearly as focused and quick, and very much
dependent on imagery. But imagery is not enough. Local knowledge is
needed to name places, and identify features … most roofs don’t tell
you what’s going on inside a building. In Haiti, two deployments have
gone to advocate, train, and build capacity for OSM locally. Nicolas
is now planning to go again, as part of our long term plan for OSM in
Haiti, to continue the work to the point where the Haitian community,
which includes their government and civil society actors, are ready to
take it forward.

More information about the Talk-au mailing list