Re: [ontology-summit] Presenting diverse data sets as linked data - hack

[Ivan Kurmanov <duraley@xxxxxxxxx>]

Victor,

So, it sounds like you want to explore a way to build flexible, adaptable (configurable, probably) interface to various kinds of data.

Like, as if we take

1. source data

2. mapping to an ontology (esp. ISO 15926-2)

3. some set of data presentation rules or configuration, that is easy to modify

and try to produce a human-friendly web interface based on all the above?

So the real challenge is going to be to design (and implement) a language for presentation rules for various data?

A set of such rules could accompany an ontology and increase the value of both the data and the ontology, i believe. Sounds interesting to me. 

-ivan

2014-02-06 Victor Agroskin <vic5784@xxxxxxxxx>:

Ivan,

> Date: Wed, 5 Feb 2014 16:29:35 +0300
> From: Ivan Kurmanov <duraley@xxxxxxxxx>
> Subject: Re: [ontology-summit] Presenting diverse data sets as linked
>         data - hackathon idea

>
> Hi Victor and everyone,
>
> So, in simple terms, your idea is to take an existing dataset or several,
> map them to the ISO15929 ontology, and make a nice human-friendly web
> interface to it?

Not so simple :-) I'm not interested in the solution of a static
problem - to take several datasets and design a nice representation
for data. Assume there appears a new dataset with partially different
data model. I need an approach and instruments which will allow me to
redesign a page look to reflect newly discovered relations. Or assume
there is a new dataset with new types of entities in it. They'll
require completely different page design - other relations deemed
essential, other classifiers seen as primary classifiers, etc.

> It could be similar to freebase.com in some ways? What
> would be the difference?

I'm not familiar with freebace.com. But at first look its data model
is extensible much like RDF model is, without restrictions. My models
do have modelling restrictions from ISO 15926-2 ontology.

But the main question is the following - is freebase concept page
content (template, look and feel) easily managed in terms of governing
ontology? Can I select relations to show, define inheritance along
some of them? Can I look for patterns and visualize only nodes linked
to my concept with chains of length 3, and so on.

If not - this is the difference. If yes - probable freebase can be a
good platform to compete in the same task.

Regards,
Victor Agroskin
TechInvestLab.ru

>
> (I'm want to take part in the hackathon and looking for a team & project to
> join.)
>
> Ivan

On Tue, Feb 4, 2014 at 2:28 PM, Victor Agroskin <vic5784@xxxxxxxxx> wrote:
>
> To bring a content of a data source into the world of Linked Data can pose a significant challenge. Especially if the data set is not uniform, has a high degree of variety and data model for it can not be easily obtained. In such cases help can come from an ontology where these data can be seen as instances.
>
> In the course of our software tool development  we became very interested to test this approach. We've started integration of our dot15926 Editor (http://techinvestlab.ru/dot15926Editor) with a web framework to allow easy and extendible definition of web page templates for presentation of complex interlinked ontology concepts.
>
> We are proposing the following hackathon scenario for teams with similar interests:
>
> 1. Select a data source which is either already connected to some ontology (some RDF data set, for example), or whose meta-data (data model, table headings, etc.) can be mapped to one.
>
> 2. Define convenient presentation form for humans in terms of ontology patterns representing domain knowledge.
>
> 3. Make software prototype for the whole data-transformation chain or a part of it:  raw data - mapping - ontology-linked data - pattern search -  Linked Data web page presentation.
>
> One of the key tasks for our team is to explore extendibility of this approach. We'd be happy to see tools which will allow step-by step development of data presentation as new and varying data sources are added to the project and new relevant patterns are identified.
>
> We are looking for:
>
> - people with data sets and relevant domain knowledge interested in bringing these data to Linked Data world (engineering data sets are a priority for us, as we're working with ISO 15926 ontology);
>
> - people ready to test our tools (Python knowledge is required to work in dot15926 Editor environment, current version supports only part of the required functionality, access to new beta version of the software will be provided);
>
> - people with their own tools and approaches interested in the scenario described above.
>
> Our team will be virtureal, working online and in our real office in  Moscow, Russia.
>
> Come on board!
>
> Victor Agroskin
> TechInvestLab.ru
>
>

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
Community Portal: http://ontolog.cim3.net/wiki/

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014  
Community Portal: http://ontolog.cim3.net/wiki/     (01)