The GeoWeb Ecosystem and the Future of Attributed Data
July 13th, 2007by Sean Gorman
Chris’s latest round of posts on the KML 2.2 specification got me thinking about the over all GeoWeb ecosystem, how it is structured and how it will evolve in the future. The term GeoWeb was coined by Mike Liebhold of the IFTF to encompass ““The integration of Web-accessible digital cartographic information with geocoded Web hypermedia.” The term and concept was picked up by several including John Hanke as the descriptor of Google’s activity in the space, which he has said will be as big as the page web.
We had a good talk with Michael Jones at Google Developer Day about the GeoWeb ecosystem and what he saw as Google’s role in it. Michael saw Google as the provider of the base layers of the ecosystem – providing the foundation for others to build upon. A set of technologies and data that enable others to create applications to solve a wide variety of problems.
Providing these technological and data layers is like providing the infrastructure to a modern economy that underpins the commerce built on top of it. It is a great business position to be in that creates very positive economies of scale, which the market naturally pushes towards having a few competitors in. Lots of historical examples you could get into at this point, but I’ll try to keep out of the weeds. Currently the GeoWeb has a couple players competing with Google namely Microsoft and Yahoo. There are some other players in the game nipping at the periphery but either do not see the bigger picture – like Mapquest being stuck in driving directions and seeing little use in Web 2.0 or user generated content – or do not have enough resources to bring the economies of scale to bear to play with the big boys.
So, what is the foundation of the GeoWeb ecosystem that Google, Yahoo and Microsoft are competing over? The beginning was packaging up three layers of data into intuitive mapping applications:
- business listings (local search)
- satellite imagery
- road data and boundaries
This was a huge success for all three and eventually they opened API’s for others to access it and folks began building applications on top of their mapping platforms. At this point the technology went from being applications to being infrastructure that others were building their own applications on top of. In addition to the base data layers and API’s to access them other services emerged like geocoding, driving directions and traffic. Then an organic base layer emerged – user generated content. Initiated by a host of start ups like Platial, Tagzania, Frappr, Flickr/GeoBloggers, Panaramio etc. First Microsoft responded with features in Virtual Earth that allowed you to create your own data on the map then Google followed shortly there after with the more evolved MyMaps. This gave both Google and Microsoft the ability to support user generated data. In Googles case this was rolled into the KML search initiative where this data and others could be found using Google technologies. Google continued to push the envelope with Mapplets which allows third party services to be connected together in the Google ecosystem. All the applications that had been built upon the Google Maps API infrastructure could be interconnected through Mapplets and shown on a common map – Google’s. Quite the clever move if a good number of the Google mashups buy in and create Mapplets.
This post started by referencing our posts of the future of KML and a design to support attributed KML through the schema tag. Why is KML, attributed data, and the schema tag important to the future of the GeoWeb? Simply put we believe the next layer in the GeoWeb is attributed feature data. So, what is attributed feature data and why is it important? All the data we’ve talked about thus far has been locations with descriptions – most commonly embodied in the form of a push pin. When you click on the push pin you get a description of what that push pin embodies – user generated content, a business listing, a photograph etc. etc. Attributed feature data is when you want to describe a location with more than just a description. A location that has a list of attributes that describes its features in a structured and parsable means. For instance Census data provides a wealth of data describing a locations like a zip code. A zip code has more than just a description, but a list of attributes from a data table that tell the users things like the population, number of people who make over $100,000, how many men over the age of 40 etc. etc. There is a wealth of attributed data that describes locations and a lot of people that would like to create new data to decribe locations that can support multiple attributes. The KML design Chris has been posting about is our proposal of how to support this type of data in the GeoWeb ecosystem.
So, why is this important for the future of GeoWeb? It is important because it allows a wealth of existing data and databases to seamlessly join the GeoWeb. Not only can existing GIS data be easily converted, but the vast wealth of data sitting in spreadsheets and databases (like Swivel) that have a geographic identifier can be seamlessly tagged to a location in a structured way. Once you have structured attributed data for a location you can perform analysis based on those numbers. Want to find the best neighborhood based on low crime, good schools, and high income – no problem you have a set of attributes to pull from to solve the problem in a number of ways. Want to pull in user generated content about that neighborhood to verify the data statistics – no problem it all runs in the same standard and system.
One of the little advertised problems with the GeoWeb is the difficulty in monetizing it. The vast majority of revenue has come from local search and despite large amounts of investment in the infrastructure the next big payday has yet to emerge. This is where MapQuest’s disdain for the GeoWeb and social media gets to the crux of the issue. Where is the value and how are you helping users make better decisions? Our conclusion being you need structured attributed data to enable the evolution to value creation. Then the user generated content has true value because it is the context of something that can be used to make a solid decision. At the end of the day we believe this is what will create the bridge between hypermedia and geodata that Liebhold and others have proselytized.
Popularity: 7% [?]






July 27th, 2007 at 3:23 pm
[...] creating descriptive data (push pins with text and pictures), what is the potential of them creating feature attributed data that provides statistics about a location like population, pollution emission levels, crime rates [...]