Re: Semantics and Data Consumption

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.