Re: Semantics and Data Consumption

Hi Laufer,

On Mar 27, 2014 11:27 PM, "Laufer" <laufer@globo.com> wrote:
>
> Ig,
>
> What I am trying to expose is that we should differentiate the ideas of
the RDF Model and Linked Data from the way Data is stored.

+1

>
> Besides that, I think we should take into account the tools that are
being used do expose Data on the Web.
>

What you mean?

> Best,
> Laufer
>
>
> 2014-03-27 5:38 GMT-03:00 Ig Ibert Bittencourt <ig.ibert@gmail.com>:
>
>> Hi Laufer,
>>
>> Thank you for your didactic e-mail. :)
>>
>> I agree that Data semantics is very important and we should definitely
try to connect our the data as much as possible to some kind of schema of
others people's data.
>>
>> As far as I understand, your proposal goes in the same way as the fifth
start of the Tim's 5 start open data plan [1] and also with the third
principle of the Linked Data Principles [2]. Is that right?
>>
>> Even though, IMHO perhaps could be a good idea to reinforce the LD
principles as best practices.
>>
>> [1] http://5stardata.info/
>> [2] http://www.w3.org/DesignIssues/LinkedData.html
>>
>> All the Best,
>> Ig
>>
>>
>> 2014-03-25 16:25 GMT-03:00 Laufer <laufer@globo.com>:
>>
>>> Hello, All,
>>>
>>>
>>>
>>> I apologize for the long message.
>>>
>>>
>>>
>>> I would like to talk about some concepts that are being discussed by
the WG and are related to Data Formats and Semantics
>>>
>>>
>>>
>>> Bernardette published a page in the wiki where she defines phases for
the Data on the Web Lifecycle.
>>>
>>>
>>>
>>> When we inspect some of the Use Cases and the Stories listed in the
wiki, including the webinars presentations, we can see that there are more
than one player, a chain of players, that is responsible for allowing the
consumption of Data.
>>>
>>>
>>>
>>> The Data Generation and the Data Distribution phases are done by
persons that access the raw data to be published but use platforms for
distribution that have their own metamodels as, for example, CKAN and
Socrata.
>>>
>>>
>>>
>>> The issue "what is the Data format that is consumed" is mixed with the
idea that the Data format of the stored Data is the same format of the
consumed Data . In some Use Cases we can see, in some instances, that the
Publishers store different formats to be downloaded by the Consumers.
>>>
>>>
>>>
>>> At first sight, it is not important what is the Data format that is
stored in the repository. When someone request Data, the transformation
(serialization) of the stored Data could (should?) be done by the Data
provider.
>>>
>>>
>>>
>>> Let's take Socrata as an example. A Dataset in Socrata could be
uploaded from an Excel file, but once it is stored in Socrata cloud, we
don't know what is the Data format of the original Excel file that is
stored as a Dataset. A Data consumer has a standard interface where she can
browse the Dataset and she can ask the platform to export Data in different
formats, including pdf, json, xml, rdf and xls.
>>>
>>> Socrata also provides an individual Endpoint with an API for each
Dataset. It considers the Endpoint as a way of exporting Data, a slice of
the whole Dataset.
>>>
>>>
>>>
>>> When we think about Data semantic, this semantic should be described as
metadata. It can be stored, for example, in a pdf file describing the data
model, in a technical style or in a free style. What is important is that
the Consumer could understand what is being said about the Data that she is
consuming.
>>>
>>>
>>>
>>> What could be a Best Practice would be to use a more wide common
understanding of this metadata. This is one of the contributions of rdf
model when it defines the use of common vocabularies as a way to describe
the properties of resources. Besides that, it also introduces the idea of
universal identifiers in a way of linking Data from different Datasets.
>>>
>>>
>>>
>>> There is a huge amount of Data to be loaded on the web that has its own
semantics. People can publish these Data in his own view letting the
developers to understand each one of these semantics and making the
mashups. It's ok. But if the Publishers could use common vocabularies these
could facilitate the work for the Developers to integrate Data.
>>>
>>>
>>>
>>> Let's take an example. In NYC Open Data Dataset "311 Service Requests
from 2010 to Present" there are two columns labeled "Latitude" and
"Longitude". The type of these two columns is Number. Well, we can guess
that they are related to the latitude and longitude of the address where a
service was requested.
>>>
>>>
>>>
>>> There is a human interface where it is possible to browse the Dataset:
>>>
>>>
https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/stnw-hdrd
>>>
>>>
>>>
>>> To get the information about a service request we can use the Endpoint
to export Data in json or rdf formats. The columns labels are identified by
property names derived form the columns labels: "Latitude" is identified as
"latitude"; "Longitude" as "longitude."
>>>
>>>
>>>
>>> Using the endpoint created for the Dataset we can obtain the json
output of the first row:
>>>
>>>
>>>
>>> http://data.cityofnewyork.us/resource/stnw-hdrd.json?$limit=1
>>>
>>> [ {
>>>
>>>
>>>
>>> "longitude" : "-73.76983198736392",
>>>
>>> "latitude" : "40.71159894212768",
>>>
>>>
>>>
>>>  }  ]
>>>
>>>
>>>
>>> Using the endpoint created for the Dataset we can obtain the rdf output
of the first row:
>>>
>>> http://data.cityofnewyork.us/resource/stnw-hdrd.rdf?$limit=1
>>>
>>>
>>>
>>> <rdf:RDF
>>>
>>> xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>>>
>>> xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
>>>
>>> xmlns:socrata="http://www.socrata.com/rdf/terms#"
>>>
>>> ...
>>>
>>> xmlns:dsbase="http://data.cityofnewyork.us/resource/"
>>>
>>> xmlns:ds="http://data.cityofnewyork.us/resource/stnw-hdrd/"
>>>
>>> xmlns:usps="http://www.w3.org/2000/10/swap/pim/usps#">
>>>
>>>
>>>
>>> <dsbase:stnw-hdrd rdf:about="
http://data.cityofnewyork.us/resource/stnw-hdrd/27702159">
>>>
>>> <socrata:rowID>7055868</socrata:rowID>
>>>
>>> <rdfs:member rdf:resource="
http://data.cityofnewyork.us/resource/stnw-hdrd"/>
>>>
>>>
>>>
>>> <ds:latitude>40.71159894212768</ds:latitude>
>>>
>>> <ds:longitude>-73.76983198736392</ds:longitude>
>>>
>>>
>>>
>>> </dsbase:stnw-hdrd>
>>>
>>> </rdf:RDF>
>>>
>>>
>>>
>>>
>>>
>>> Well, the rdf does not introduces any kind of semantics in this case.
It is only a different serialized format of the Data returned in json. The
property http://data.cityofnewyork.us/resource/stnw-hdrd/latitude doesn't
have more semantics than the label "Latitude".
>>>
>>>
>>>
>>> But Socrata allows the owner of the Dataset to associate an rdf
property to a column. The user can associate any URL as a metadata of the
column and, besides that, Socrata lists some properties that it understands
from some vocabularies: dcat; foaf; dublic core; geo.
>>>
>>>
>>>
>>> I associate to the column "Latitude" the URL:
http://www.w3.org/2003/01/geo/wgs84_pos#lat
>>>
>>>
>>>
>>> I associate to the column "Longitude" the URL:
http://www.w3.org/2003/01/geo/wgs84_pos#long
>>>
>>>
>>>
>>> I made the endpoint call again:
>>>
>>>
>>>
>>> http://data.cityofnewyork.us/resource/stnw-hdrd.rdf?$limit=1
>>>
>>> <rdf:RDF
>>>
>>> xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>>>
>>> xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
>>>
>>> xmlns:socrata="http://www.socrata.com/rdf/terms#"
>>>
>>> ...
>>>
>>> xmlns:geo="http://www.w3.org/2003/01/geo/wgs84_pos#"
>>>
>>> ...
>>>
>>> xmlns:dsbase="http://data.cityofnewyork.us/resource/"
>>>
>>> xmlns:ds="http://data.cityofnewyork.us/resource/stnw-hdrd/"
>>>
>>> xmlns:usps="http://www.w3.org/2000/10/swap/pim/usps#">
>>>
>>>
>>>
>>> <dsbase:stnw-hdrd rdf:about="
http://data.cityofnewyork.us/resource/stnw-hdrd/27702159">
>>>
>>> <socrata:rowID>7055868</socrata:rowID>
>>>
>>> <rdfs:member rdf:resource="
http://data.cityofnewyork.us/resource/stnw-hdrd"/>
>>>
>>>
>>>
>>> <geo:lat>40.71159894212768</geo:lat>
>>>
>>> <geo:long>-73.76983198736392</geo:long>
>>>
>>>
>>>
>>> </dsbase:stnw-hdrd>
>>>
>>> </rdf:RDF>
>>>
>>>
>>>
>>>
>>>
>>> Well, the rdf returned geo:lat and geo:long as the properties of two
numbers that has a well known semantics.
>>>
>>>
>>>
>>> For me, this is a Best Practice.
>>>
>>>
>>>
>>> What do you think about this?
>>>
>>>
>>>
>>> I apologize, again, for the long message.
>>>
>>>
>>>
>>> Kind Regards,
>>>
>>> Laufer
>>>
>>>
>>>
>>> --
>>> .  .  .  .. .  .
>>> .        .   . ..
>>> .     ..       .
>>
>>
>>
>>
>> --
>>
>> Ig Ibert Bittencourt
>> Professor Adjunto III - Universidade Federal de Alagoas (UFAL)
>> Vice-Coordenador da Comissão Especial de Informática na Educação
>> Líder do Centro de Excelência em Tecnologias Sociais
>> Co-fundador da Startup MeuTutor Soluções Educacionais LTDA.
>
>
>
>
> --
> .  .  .  .. .  .
> .        .   . ..
> .     ..       .