Comparing OSM and Google Services in Kenya and the Developing World

After reading Erik’s post on Google driving directions in Kenya, my thoughts wandered to how services built on OpenStreetMap data compared in Nairobi and Kenya. The main difference I can see is that with OSM services, when something is deployed anywhere, it’s usually immediately available globally, so we aren’t left waiting for opaque corporate processes to gift us with new features.


I’ve compared MapMaker and OSM coverage before, and found them to be nearly equivalent in areas with high resolution satellite imagery. Yahoo has less satellite coverage generally, over a smaller area of Nairobi, so that is where you see the highest concentration of OSM data. Outside Nairobi, OSM relies mostly on the FAO Africover import, with select places surveyed in more detail .. I think mainly vacation spots :). This is quickly being supplemented as Map Kibera folks are borrowing GPS for their travels up countries in the festive season.

Both Google and Yahoo imagery are over four years out of date. Anyone familiar with Nairobi’s rapid recent building spree can see it clearly from comparison with satellite imagery with known timestamps. This means that provider’s satellite imagery alone is not sufficient to map here .. you need both up to date imagery, and in situ surveying. That’s where GPS and Walking Papers show their strength in data collection.


OSM has a very active developer community focused on routing. The products aren’t quite a slick as Google’s offerings, but just as powerful, mostly based on pgRouting. CloudMade’s routing is based on adding pins to the map, rather than search, but otherwise do a comparable job to Google’s routing choices.


Like Google, CloudMade routing lacks traffic data in Nairobi. I do know that there are folks in Nairobi working on deploying traffic sensor systems. And folks working on matatu routes. Now t>he key thing is how we will see their data in maps. They could negotiate with Google to have their data included, and I can only wish them luck and a prayer for something like a good deal. But there is no need to wait for Google bureaucracy to start helping improve Nairobi traffic. They could simply build their own routing application with open source data and tools, that integrates their traffic sensor network.


Erik seems to be having fun playing with his iPhone in Nairobi 🙂 (I stick to my solar powered, mPesa enabled phone here). Not many folks have iPhones in Kenya yet .. though you can find Chinese knockoffs on streetside mobile kiosks downtown. There’s no iPhone OSM routing yet. Still, there are a couple apps which offer really key features for Nairobi, and I hope Erik finds a chance to give these apps thorough testing here too.

OffMaps is local caching of maps on your iPhone, which means you can store all of Nairobi locally with no need to spend buckets of airtime repeatedly downloading maps.

MapZen POI Editor is collection of OpenStreetMap points of interest on your iPhone. It’s probably the most user friendly way to contribute to OSM. Now I don’t think it’s entirely fair to Google that Erik critiques the misplacement of his father’s office on Upper Hill. The whole idea with collaborative cartography is that the map can be improved by anyone. However, with OSM or MapMaker, you usually need to keep notes on mistakes you see in your business (I have several of these filled with corrections). MapZen allows this to happen right there in situ, as you see the errors on the street you can immediately correct them.

The Point

Now I wouldn’t mind buying drinks for Google employees. Now, most of the folks hired by Google to fill in data on MapMaker aren’t working there any more, so they may appreciate the drink more than ever (just kidding guys!).

Yes, I did say “hired by Google”. Though they claim to be working within a community, the overwhelming contributor is Google themselves. What percentage is internal or external to Google, I don’t know, because they don’t release the data to calculate those sort of stats. For OSM, we can plainly see which individual contributed how much, and produce all manner of stats. Though OSM has jsut a few folks producing the majority of data, that curve is flattening out rapidly.

It’s great that Google is extending it’s services in Kenya and the developing world. Heck they even have a bus in India (we’re working a mapping matatu here ;). But the point is that with open source and open data, people everywhere don’t have to wait for Santa Google to gift them with new features .. all the tools are readily available for maps to leap frog in the developing world even more than the mobile phone.

7 thoughts on “Comparing OSM and Google Services in Kenya and the Developing World”

  1. Nice piece Mikel, I myself suspect that some of these Google employees have collected map data “under a tree” as we say in my business: When comparing Google and the meager OSM in 2008 when I landed I was very impressed with Google – so precise, so massive. However, I quickly found that a lot of places had the same mistakes as roadmaps of Nairobi online…

  2. A Kenya Buzz ( had really cool Matatu route visulisation done in summer 2009, unfortunately this has not been opened up and shared with the masses. I heard Phares Kaboro (@kaboro) mentioning something about a proposed traffic monitoring solution sometime ago, I do not know how that has faired.

  3. majiwater .. yes, that Matatu route visualization is incredible but totally locked down. We’ll do better. And Phares is exactly who first told me about traffic monitoring; there other efforts as well. Soon enough…

    Bo .. do you have any examples of that?!

    Jorge .. thanks, will recommend that too…

  4. Sorry, not at this point. The obvious ones I saw in 2008 are gone now. Will have a look over Christmas though.

Comments are closed.