Boris Mann

Open Source. Community. Decentralized Web. Building dev tools at Fission. Cooks & eats.


Semantic Web Community Barn Raising in Vancouver

Or, how I got a tasted of Linked Data and wanted more.

I've been noodling about with a variety of semantic web / open data concepts for some time. Most recently, I spoke at Drupalcon DC 2009 on Practical Semantic Web (and why you should care). There's a video of the presentation embedded in that last link, and the presentation is available on SlideShare.

I think I can explain things even more simply: we can place a simple piece of data on a web page - like an address, or marking something as a person. From those source pieces of data, we can link it to ever more pieces of information. Except, instead of just "bare links", these links carry meaning, creating a richer web of content. In my presentation, I used the phrase "RDF is food for robots" - a way to share more info about the visible, human readable information that is already there.

From my point of view, that's already cool enough. But it's still not very *practical*. Why should you care, today? That leads me to open data, in general. The Dev Seed guys in DC have a great example of this in their StumbleSafely site -- it mashes up bars in Washington DC with crime statistics, cross-referenced with time for day, night, and evening data. screenshot

We don't have anything like this in Vancouver. What we do have is some interesting data that is beginning to be gathered in Freebase, and more specifically in the Vancouver-specific corner of it, known as vanbase. Looking at the StumbleSafely application, and thinking about some data that I hate to see duplicated over and over again, I came up with the idea of restaurants, and extra information about them.

The concept: patios, beer, and breakfast. Wouldn't you like to know where the best patios, beer, and breakfast restaurants are in Vancouver? So, in the morning, the site will show strips of bacon highlighting breakfast places. As it gets later in the day, umbrellas will pop open on the map, showcasing patios. Later at night, as the site goes dark, glowing beer steins appear, showcasing great beers on tap around the city. Through linked data, we can go on to surf the mass of information about brands of beer.

There are dozens of restaurant sites, in various stages of overlap. None of them share data, and they're not really about data in any case -- they succeed or fail because of SEO, or because of user features on their sites. And anyone can get the data -- pay a couple of thousand and you can get all the restaurants across Canada.

Linked Open Data is all about having shared information sources and building value on top of them, through those linkages.

The concept will be to fill in vanbase with restaurant information, starting with this list of Vancouver restaurants. Then we'll work with a Drupal (and possibly other) front end to do display of this rich semantic web data in different interesting ways. Why Drupal? It's got a really great set of semantic web tech already implemented, and can then show how an out of the box CMS can participate in a web of linked data. But hopefully (since this is also about open data) we can get other front ends to the data, or even have some of these existing restaurant sites contribute their data sets to Freebase.

Now, all that is quite a lot to think about and do. I started the title of this post by calling it a Semantic Web Community Barn Raising - I hope to see many people locally get involved in everything from data entry, to development, to design in building out a semantic data set, and then showcasing what we can do with it.

So, I'm calling an initial brainstorming meeting this Thursday morning over breakfast, 8:30am at Deacon's Corner. Here's an event entry on UQ Events to help you remember. @notmatt, @dale42, and @jimpick will be there, hope to see some of you as well. For more info, check out the Google Site SemWeb Restaurants, where we'll be posting notes and ideas as the project progresses.