[Data-modeling] Importing Wikipedia infoboxes

Robert Cook robert at metaweb.com
Thu Oct 16 03:27:23 UTC 2008


Hi Alf -

Where wikipedia templates (of which infoboxes are a subset) have  
values that are links to other wikipedia articles or where the values  
are clean dates or numerical literals (with or without units), we can  
extract data.  Interestingly, your first example is entirely  
extractable with our system, whereas the second isn't because it  
mostly uses literal strings that don't refer to other wikipedia  
articles.

If you let us know which Freebase properties you'd like to hold these  
values, we could do that mapping and load the information.  In the  
case of many infoboxes, there are no existing Freebase properties, but  
if you suggest them, or, better yet, create the properties and types  
in a private domain, we'd be happy to get the data loaded.

Robert

On Oct 15, 2008, at 4:01 PM, Alf Eaton wrote:

> My recent plans to make some saved views in Freebase have been
> scuppered by Wikipedia infobox data being missing from the imported
> items. Cities, for example:
> http://www.freebase.com/view/en/winchester
> vs
> http://en.wikipedia.org/wiki/Winchester
>
> or theatres:
> http://www.freebase.com/view/en/almeida_theatre
> vs
> http://en.wikipedia.org/wiki/Almeida_Theatre
>
> Does anyone know what needs to happen so that more of the infobox
> information can be imported into Freebase?
>
> alf
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling



More information about the Data-modeling mailing list