W3C home > Mailing lists > Public > public-dwbp-wg@w3.org > March 2014

Re: Semantics and Data Consumption

From: Phil Archer <phila@w3.org>
Date: Fri, 28 Mar 2014 12:22:00 +0000
Message-ID: <53356968.3030501@w3.org>
To: Laufer <laufer@globo.com>, Ig Ibert Bittencourt <ig.ibert@gmail.com>
CC: Public DWBP WG <public-dwbp-wg@w3.org>


On 27/03/2014 21:31, Laufer wrote:
> Ig,
>
> I think it is important to inspect tools like Socrata and DKAN that are
> used to Publish Data. I like the term, exposing Data.
>
> When you see a Use Case as dados.gov.br what you have is a portal using
> DKAN as a tool to expose Data. The same thing with NYC OpenData or
> ControlPanel LA, but using Socrata. They have a way to manage catalogs but
> the real users are different agencies that constitute other kind of Use
> Cases.
>
> DKAN and Socrata have their own metamodels. These metamodels define the
> things. They define the way Data is exposed when you use the tool.

That's where we I believe should have something to say - how data is 
exposed. How it is stored internally is irrelevant. I believe CKAN does 
no more than simply keep uploaded files as they are, it doesn't create 
in internal copy that is somehow stores as part if its own structure.

>
> I think it would be interesting to have people from these organizations as
> participants of the WG.

Socrata are W3C members and I have repeatedly invited them to join this 
and the CSV WG, so far without success. But I will keep trying.

  They could tell us how they see this market and at
> the same time they could talk about their metamodels and they could be one
> of the targets of the Best Practices, including in their metamodels
> features that could help to implement the recommendations of the WG.

In theory, CKAN exposes metadata about the datasets using DCAT. They 
don't do it very well, however. If there is one question I can be sure 
I'm going to be asked at pretty well every event I go to it's "why is 
CKAN so bad at using DCAT?" As you can imagine, I talk to OKFN about 
that from time to time ;-) Their model remains focussed on helping 
humans find the data rather than machines. However, the European 
Commission's DCAT Application Profile [1] is, I hope leading to change. 
Since the EC uses CKAN and wants its own portals, and everyone else's, 
to use DCAT in the same way, there is pressure and, more importantly I 
believe money, forthcoming to fox this in CKAN.

On the more general point - yes both Socrata and, increasingly, CKAN, 
offer APIs to data. This is all focussed, understandably, on CSVs. 
Nothing wrong with that, but it's just an API to extract that data. What 
I hope the CSV on the Web WG will lead to is the portals supporting use 
of that new standard in due course. Rufus Pollock (OKFN's CEO) is in the 
CSVW WG.


Phil.

[1] 
https://joinup.ec.europa.eu/system/files/project/DCAT-AP_Final_v1.00.html

>
> Best Regards,
> Laufer
>
>
> 2014-03-27 18:11 GMT-03:00 Ig Ibert Bittencourt <ig.ibert@gmail.com>:
>
>> 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.
>>>
>>>
>>>
>>>
>>> --
>>> .  .  .  .. .  .
>>> .        .   . ..
>>> .     ..       .
>>
>
>
>

-- 


Phil Archer
W3C Data Activity Lead
http://www.w3.org/2013/data/

http://philarcher.org
+44 (0)7887 767755
@philarcher1
Received on Friday, 28 March 2014 12:21:31 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:24:12 UTC