Statistics

[Tom Grahame]

I was unsure how to model the statistics class (my fault for not being at the appropriate meeting) so I've made two attempts:

https://gist.github.com/tfgrahame/7489281

A type hierarchy that I am not keen on due to the large number of classes this creates. The only advantage I can see with this approach is enabling a statistic to have more than one property, for example subScore with both period and score values.

https://gist.github.com/tfgrahame/7531237

Much more concise with statistics represented as properties as values. I much prefer this option.

I have only modelled soccer thus far as an illustration of the different approaches. Although I know how schema.org works in principle, I don't know it well enough to say how either approach would fit in to its model.

I'd really appreciate any help or opinion as to how to solve this.

Tom

[Vicki Tardif Holland]

I prefer the second, more concise schema. It may be easy to get wrapped around the axle trying to come up with all of the possible properties needed to describe every sport in full. I think we should focus on a basic, common set for the general description of a sporting event and then iterate to model each sport individually.

- Vicki

Vicki Tardif Holland | Ontologist | vta...@google.com 

 

--
You received this message because you are subscribed to the Google Groups "sports-schema-collab" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sports-schema-co...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Jason Johnson]

Agreed.  I think we should move forward with second approach.

To unsubscribe from this group and stop receiving emails from it, send an email to sports-schema-collab+unsub...@googlegroups.com.