I have to construct a dataset of hospitals. For each hospital, I will indicate the the town/region in which it is located.

I already have dataset of towns/regions of the concerned geographic area.

In order to not repeat informations about towns/region in my new hospital dataset, I wonder if is it possible to be linked to my existing towns/regions dataset.

Many Thanks.

  • Linked Data is an approach designed specially to solve this issue. Linked CSV is a set of conventions for CSV files that enable them to be linked together. I can't understand how your question is related to data.gov, probably these approaches is inapplicable in your case. Jul 28 '17 at 7:09
  • Thanks @StanislavKralin. Please, I want to know why Linked Data is not applicable is not applicable in my case ?
    – Moh_BOB
    Jul 28 '17 at 13:21
  • It sounds like it could be useful to make your data suitable for putting into GIS by giving each hospital co-ordinate data (perhaps it already comes with geographic location?) Jul 28 '17 at 18:16
  • Yes, each hospital has co-ordinate data. But the administrative location (town/department) is as important as co-ordinate. Thanks
    – Moh_BOB
    Jul 29 '17 at 8:45

First model the relationships from the geographic area and probably expand it more:

Dallas > located_in > Texas
Dallas > country > USA
Dallas > located_in > USA

Texas > first_level_administrative_division_of > USA
Texas > located_in > USA

Then do the same for the hospitals list, and merge those that match completely.

You can also create a fake identifier of the kind USA/Texas/Dallas and again merge on that criteria. The tree can be as deep as you want.

Hope this helps.

  • Thanks @marfi. Please, I want to know if the data.gov of countries are constructed in a linked way ?
    – Moh_BOB
    Jul 28 '17 at 13:21
  • data.gov is a portal, what file in particular you are working with?
    – marfi
    Jul 29 '17 at 14:42
  • Sorry, I didn't know that "data.gov" was the US open data portal. By data.gov I meant to say a generic open data of a country.
    – Moh_BOB
    Jul 29 '17 at 15:30
  • It is not getting any clearer. You have to give a specific example. In Europe we use LAU and NUTS in US you use FIPS, it is not one standard.
    – marfi
    Jul 30 '17 at 11:22

Why don't you assign a ID to each relevant geographic combination in your address database:

ID; Town; State; Country

1 ; Miami; Florida; USA

after you did than you could add a column to your hospital list (like "geo_id") and assign the correct geo_id to each hospital.

hospital name ;[...]; geo id

as a result the geo_id would link to the correct information on the other table.

If you have to join both tables you can do that by "hospital.geo_id = address.id"

  • 1
    Ok, your idea is the basic of relational databases. But I want do this in Linked Data manner. The difference is that in the later, this contribute to the web of data.
    – Moh_BOB
    Aug 1 '17 at 11:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.