[Data-modeling] indian elections

Kirrily Robert kirrily at metaweb.com
Thu Apr 9 21:44:27 UTC 2009


Hi Chris,

On 09/04/2009, at 2:00 PM, Chris Blow wrote:
> I am working on a project that seeks to organize information about the
> upcoming Indian Elections.
>
...

> I have been recommending that we use Freebase to structure some of our
> data. Can anyone give any advice about their experience modeling
> elections?


My advice would be to start out building a base for your project.   
 From there, see if you can extract the stuff you need as saved views  
against our Government Commons.  For instance, for candidates in this  
election, create a saved view on /government/politician using the  
"election campaigns" property to connect to the election in  
question.   Repeat for each of the points you listed, other than the  
violations of the ECMCC which we don't have any modelling for  
whatsoever (see below.)

As you create each view, look at whether we had the structure you  
needed.  (We may not have the data -- our coverage for India is not  
great -- but I'm talking more about the availability of properties/ 
fields that cover what you need.)  If we are missing fields that you  
need, and which you feel would be beneficial for the Freebase Commons,  
please feel free to ask for them either on this mailing list, or in  
discussion posts on the Government Commons on Freebase.com.

If you find that you need fields/properties that are specific to your  
own project, you can create your own schema in your base to do that.   
The best thing to do is to create types based on (i.e. including)  
existing Commons types, and add what properties you need.  I imagine  
the violations you mention will have to be modelled from scratch in  
this way, as we don't currently have any schema like that, and it  
seems difficult to generalise.  Again, ask here and/or on the  
discussion boards if you need help with anything specific.

K.

-- 
Kirrily Robert
Freebase Community Director
kirrily at metaweb.com





More information about the Data-modeling mailing list