Jim Ley pointed me in the direction of ChefMoz, a collaborative restaurant guide organised by DMoz. It looks like a contemporary of DMoz and is run by Netscape but how come I've never heard of it? Obviously I don't go out enough...

The best thing though is that,like its better-known sibling, ChefMoz provides RDF dumps of its database although Jim tells me that he had problems parsing it in the past. I suspect it might be character encoding issues since there appears to be some non-ascii characters in there.

The main RDF file is about 100MB decompressed and contains nearly 400,000 restaurants.

The restaurants are characterised by address, price, cuisine, parking facilities, credit cards accepted, dress code and whether smoking is allowed. They can also be rated on food, service and abience with an overall score as well.

The address is broken up into address, neighbourhood, city, state, country and zip components, obviously an American slant but then my version is dictinctly British :). There's an interesting element CrossStreet which is The nearest street that crosses the street that the restaurant is on. An interesting idea which could be expanded: nearest subway station, tram stop or landmark etc. There is some evidence of this in the data:

<CrossStreet>between St Marys Cathedral and Swimming Pool</CrossStreet>

There is some attempt at capturing opening hours. There is an element called Hours which contains a human-readable description of the opening times, and a ParsedHours element which represents a machine-readable version: the hours are converted to 24hr clock and split into the seven days of the week, delimited by pipes:

<Hours>Monday to Sunday 11:30am - 10:30pm</Hours>

I couldn't see any documentation that tells me what they consider to be the first day of the week but here's a more complex example which leads me to conclude that the first day is Sunday:

<Hours>Monday to Friday 12:00noon - 2:30pm, 6:00pm - 10:00pm; Saturday, Sunday, Public Holiday 6:00pm - 10:00pm</Hours>


Other posts tagged as rdf, technology

Earlier Posts