Loading ...
Sorry, an error occurred while loading the content.

[ai-geostats] Missing values and multiple objects on single location

Expand Messages
  • Bernardo Alves Furtado
    Dear List Members,   I have worked a little with GEODA and I am now working on a real estate price project with nearly 3000 observation spread throughout the
    Message 1 of 2 , Dec 12, 2005
    • 0 Attachment

      Dear List Members,

       

      I have worked a little with GEODA and I am now working on a real estate price project with nearly 3000 observation spread throughout the city. I have their location based on neighborhoods, so there are many observations for the same polygon. My questions, if you can please help, are:

       

      1)      ArcGis does not match multiple database DBF with polygons! Does it?

      2)      Mapinfo does. However, all the missing values I had been using in stata for instance, become 0s in Mapinfo which will compromise any regressions I may do. Upon this I tried to set the fields as character, but then, of course, GEODA does not recognize those as numbers and therefore cannot proceed with the regression.

      3)      So, I guess, what I want is to use MAPINFO to make the association of the data with the polygons, but maintaining those missing values as missing.

       

      I hope you guys can help.

      Thanks a lot.

      Bernardo Furtado



      Doutorando em Economia
      Cedeplar - Centro de Desenvolvimento e Planejamento Regional
      Faculdade de Ciências Econômicas, UFMG (http://www.cedeplar.ufmg.br)


      --
      Esta mensagem foi verificada pelo sistema de antivírus e
      acredita-se estar livre de perigo.
    • Roger Bivand
      ... I m assuming that you know which polygon each house belongs to. I can t answer for ArcGIS or Mapinfo. The problem is, I think, not to associate the houses
      Message 2 of 2 , Dec 12, 2005
      • 0 Attachment
        On Mon, 12 Dec 2005, Bernardo Alves Furtado wrote:

        >
        > Dear List Members,
        >
        >  
        >
        > I have worked a little with GEODA and I am now working on a real estate
        > price project with nearly 3000 observation spread throughout the city. I
        > have their location based on neighborhoods, so there are many
        > observations for the same polygon. My questions, if you can please help,
        > are:
        >
        >  
        >
        > 1)      ArcGis does not match multiple database DBF with polygons! Does
        > it?
        >
        > 2)      Mapinfo does. However, all the missing values I had been using in
        > stata for instance, become 0s in Mapinfo which will compromise any
        > regressions I may do. Upon this I tried to set the fields as character,
        > but then, of course, GEODA does not recognize those as numbers and
        > therefore cannot proceed with the regression.
        >
        > 3)      So, I guess, what I want is to use MAPINFO to make the
        > association of the data with the polygons, but maintaining those missing
        > values as missing.
        >

        I'm assuming that you know which polygon each house belongs to.

        I can't answer for ArcGIS or Mapinfo. The problem is, I think, not to
        associate the houses with the polygons as an aggregate, but to use the
        known locations of the polygons as proxies for the unknown locations of
        the houses. In the spdep package for R, just such a scenario is provided
        in the nb2blocknb() function. An online description of the function is at:

        http://finzi.psych.upenn.edu/R/library/spdep/html/nb2blocknb.html

        If you have R installed, (www.p-project.org), try example(nb2blocknb) to
        see if this looks like your case - here the Boston housing data set is
        506 census tracts blocked into 92 townships as though the coordinates of
        the census tracts were unknown.

        Given a neighbour list for polygons, the function generates a neighbour
        list (such as that needed by GEODA) for the houses, based on their
        belonging to polygons - they end up being neighbours of all the houses of
        their own polygon and of the houses in the polygons defined as neighbours
        of their polygons. The write.nb.gal() function can be used to write out
        the neighbour list.

        You can read shapefiles into R using functions in the maptools package,
        and make the first neighbour list with poly2nb() in spdep if that
        definition of neighbours suits you, or in some other way.

        Roger

        >  
        >
        > I hope you guys can help.
        >
        > Thanks a lot.
        >
        > Bernardo Furtado
        >
        >
        >
        > Doutorando em Economia
        > Cedeplar - Centro de Desenvolvimento e Planejamento Regional
        > Faculdade de Ciências Econômicas, UFMG (http://www.cedeplar.ufmg.br)
        >
        >
        > --
        > Esta mensagem foi verificada pelo sistema de antivírus e
        > acredita-se estar livre de perigo.
        >

        --
        Roger Bivand
        Economic Geography Section, Department of Economics, Norwegian School of
        Economics and Business Administration, Helleveien 30, N-5045 Bergen,
        Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
        e-mail: Roger.Bivand@...
      Your message has been successfully submitted and would be delivered to recipients shortly.