The Problem: weak semantics

Tagging is still not a very popular task, especially in corporate environments. Many users don´t see the benefit of creating metadata to describe the actual content. A typical counter-argument to social tagging is that there are too many words for the same thing. “Even if I am tagging very hard my colleagues won´t find necessarily my pages  because they will use different words to search for the content. I don´t have enough time to insert ‘New York City’, ‘NYC’, ‘Big Apple’ etc. as labels”.

The result: Tagging facilities of enterprise software platforms like Confluence are rarely used and don´t help to index content at all. Search is mostly based on classical full-text indexing. Semantic search as seen more and more on the WWW has still not entered the enterprise realm.

The Solution: thesaurus based indexing

W3C´s Semantic Web technology stack provides means to define controlled vocabularies like thesauri which results into more and more tools and data which make use of standards like SKOS. Tagging based on thesauri means that concepts are attached to pages & documents rather than putting labels on them. Labels like ‘New York City’, ‘NYC’ and ‘Big Apple’ refer to the same concept, thus it should be sufficient if one of the various terms is used for labeling, all the other names of this certain concept should be attached automatically.

PoolParty PowerTagging is able to analyse each Confluence page and to insert concepts from a thesaurus and all of their names automatically. Users can curate all suggested tags or they can also index their spaces automically resulting in a semantic index which makes search more comfortable than ever before.

Usage: enhanced collaboration with enterprise knowledge models

There are two main application scenarios which can be realised on top of Confluence and its PowerTagging extension:

• Semantic Search: Fully integrated with Confluence´s built-in Lucene based search facility, users no longer have to type in search phrases literally: Even if only ‘New York City’ is mentioned on a page on a word-by-word basis, it´s sufficient to search for ‘Big Apple’ or ‘NYC’ and results will be generated. This feature is especially interesting for domains in which a lot of technical terms or abbreviations are commonly used or for enterprises in multi-lingual environments.
• Content recommendation: Identifying similar and semantically matching contents especially in larger Confluence instances is a crucial task: Imagine you´re working for a recruiting company and you would like to match a new open position with all people in your applicant database. Or: Imagine you´re working on technical documentation and you can provide your customers automatically with further readings. Or: Imagine you´re working on a slidedeck and you´ll see instantly if some of your colleagues have worked on similar issues recently.

Don´t re-invent the wheel again and again. Save time and money. PPP will help to fulfill these tasks when creating rich contents more efficiently than ever before. You can link similar contents within Confluence automatically and you can fetch further readings even from the WWW like from Wikipedia.

If you are interested in trying out PowerTagging, please drop us a note and we will be happy to support you!

I had at least five scenarios in mind in which SKOS and linked data in general can be combined. Take a look at the slides or watch the video to find out …

• how to publish SKOS thesauri as linked data
• how to generate SKOS from LOD sources like DBpedia
• how to make use of SKOS thesauri for entity extraction & content enrichment from LOD sources
• how to use linked data mechanisms for collaborative thesaurus management
• how to use SKOS for linked data alignment & better disambiguation

reegle.info acts as a unique clean energy information portal, targeting specific stakeholders including governments, project developers, businesses, financiers, NGOs, academia, international organizations and civil society. Alongside comprehensive country energy profiles, energy statistics and a directory of relevant stakeholders it also offers the clean energy search and an extensive glossary. There is also an insightful clean energy blog with interesting and up-to-date background information.

As reegle.info provides relevant clean energy data from several key energy open data sources as for instance OpenEI, World Bank Data or the UK Open Data Portal the reegle.info Information Gateway has a strong need for efficient and automated data management mechanisms and technologies! Therefore REEEP (The Renewable Energy and Energy Efficiency Partnership) decided to use Linked Open Data (LOD).

Linked Open Data provides a powerful way for reegle.info for sustainable data management and data integration – thereby the current reegle.info linked (open) energy data cloud came into being and looks as follows:

The figure of the reegle.info linked (open) energy data cloud above shows the model behind the scenes of the reegle.info clean energy information gateway providing an insight about sources and respective connections / links between the several sources and data sets.

For the realisation of the Linked Open Data based reegle.info system the following software components are in use:

• PoolParty Thesaurus / Vocabulary Management System
• SWCs Linked Data Manager
• OpenLink Softwares’ Virtuoso Triple Store

Going this direction the reegle.info clean energy key portal is very flexible for future expansions in the fields of data integhration and data management by new data sets from several data sources!

By the way – reegle.info is very open too – thereby the whole REEEP generated data is available via a Sparql endpoint for free re-use under the UK Open Government Data license on the reegle.info data portal!

Try it out and make use of free high quality clean energy data!

The afternoon started with the discussion of WP9 the “Open Government Data” use case. First Uroš Milošević from Institut Mihajlo Pupin (IMP) reported about the Serbian CKAN project already holding some data from the Statistical Office of Serbia. Also tools from the LOD2 stack have been and will be used for this project. Sounds great!

Then Irina Bolychevsky of OKFN continued the session announcing that a better integration between CKAN and LOD2Stack should be made  to get more RDF in publicdata.eu. Good idea! We were collecting ideas for integration and talked about e.g. a wizard for generating RDF from .csv files (ULEI is working on something like that). Also a integration of google refine has been discussed. The consortium decided to make an extraction sprint transforming a (to be defined) number of  interesting data sets from CKAN to RDF.

Finally we had a discussion if linked data is a (the) solution for CKAN to find data and find related data etc. Well i think the people in the consortium are pretty sure it is (not so sure if people from OKFN are). Irina and Mark from OKFN invited everyone to provide input to the Use Case.

This session ended with a presentation about WP9a from Jindřich Mynarz from UEP and Martin Nečaský from CU. They are developing a distributed market place for public contracts. A ontology for public contracts has been developed and is open for review on google code. Next step here will be a web application for filing/creating public contracts in RDF as linked data using tools from the stack. So all in all pretty good progress in WP9.

The third day and the plenary ended with Martin Kaltenböck from SWC and Sören Auer our project lead from ULEI presenting WP10-11-12 Dissemination, Exploitation and Project Management. First we voted for our next plenary to be in Cambridge (hosted by OKFN). Past dissemination activities have already been presented on day one, so Martin reminded us all to write blog posts about all the great things we are doing in LOD2. Next big dissemination activity and also a good opportunity to meet people from the consortium will be the European Data Forum from June 6-7 in Copenhagen.

And that was pretty much it. I as i hope all the others enjoyed three days with a bunch of great people from all over Europe working on a great project. As always it was intense but it was also fun. Hope everyone had a save trip home!

“Linked Data and SKOS. Connecting the dots.”
(Wed, Apr 18, 2012 5:00 PM – 6:00 PM CEST)

Learn more about:

• How the creation of thesauri can become more efficient when built upon existing linked data sources
• How SKOS thesauri can be aligned with LOD sources and can be published as an LOD source itself
• How linked data mechanisms can be used to improve decentralized vocabulary management
• How SKOS and linked data alignment can be used for efficient schema mapping and value mapping
• How SKOS thesauri can be enriched with linked data to realise semantic search engines in a very efficient way
• How collaborative platforms like Sharepoint or Confluence can benefit from SKOS based knowledge models in combination with linked data

This webinar is powered by LOD2.

• Make knowledge explicit and available throughout your department, your organization, your partner network, the whole world!
• Conquer the Babylonian language confusion!
• Help to make search and research activities more efficient!
• Link all your information sources in a meaningful and standardized way!
• Make skills, interests and knowledge of experts visible!

The paper (PDF) should help to make the world of thesauri and semantic knowledge models accessible intuitively.

José Manuel Alonso has been working for W3C and CTIC in many open data projects. At the Web Foundation he promotes and supports (linked) open data in developing countries. Martin Kaltenböck from SWC talked with José about ongoing activities in the area of Open Government Data.

Open Data is a powerful worldwide movement these days. Regarding open data projects in developing countries and in high industrialised countries (Europe, US, Australia et al) where do you see the main differences – regarding organisational – cultural – technical issues?

We conducted feasibility studies in Ghana and Chile several months ago, are supporting the Ghana government on the development of its national initiative and have visited and have engaged in Open Data discussions with many other countries in Africa, Latin America and Asia.
The situations are quite diverse and can vary significantly from country to country. It is always difficult to generalize, but I think there are a few important differences that can be highlighted (in no particular order):

• The amount of information available in digital form is generally much lower
• The IT infrastructure is yet to be fully developed or under development
• The capacities on the government and civil society side have to be improved
• The mobile phone is the main device to access information but data connectivity is still scarce, only available in the big cities and not at all in the rural areas
• Digital literacy related issues have to be seriously considered and addressed
• Multilingualism is an important factor, as there are dozens of dialects being spoken in many countries

Said all of the above, I would say that there are also quite a number of commonalities such as privacy and security concerns, the resistance to change but also the existence of champions within government, and the interest and willingness in civil society, that is already producing a number of interesting applications.

You are also very familiar with the concept of Linked Open Data (LOD) – where do you see the main benefit in using LOD – where do you think are the main challenges – where the main obstacles?

Having managed a few projects achieving 5-star open data, I’ve learned a thing or two about the pros and cons. I’ve been saying consistently that there are a few important issues:

• There is still little knowledge about LOD out there and it is perceived as too complex
• The demand for LOD is, hence, very low
• The tooling is not powerful enough yet, specially when compared to XML tooling and others
• The modeling part is very tough

People are used to work with XML and Web Services and believe that anything along this line such as REST+JSON fulfils most expectations and needs. But this is not fully true. In my opinion, the power of LOD resides on the linking part more than anything else. Combination of data from disparate sources using RESTful techniques is much more difficult while it’s a natural fit for LOD.
My experience tells me that for dealing with few and simple datasets, investing in LOD is not really needed, but if you want to scale up and, specially, if you want to link and integrate, then you should consider LOD. It is generally a bigger investment but it pays back for interlinking big volumes of information, facilitates re-use in multiple formats, and can get very powerful when using SPARQL appropriately as it allows access to the whole underlying knowledge base.

Where do you see the main differences regarding effort of publishing and benefit in re-use (or the re-use itself) between Open Data and Linked Open Data?

I would say that the main difference here is between using the Web as an archive for files and using the full potential of the Web. If one publishes hundreds of spreadsheets on the Web using an open format and license, he is already doing Open Data, but more than using the Web, he is going back to the FTP days. And that is not too different from giving away a USB stick with the files. We can do much better nowadays.

The often cited Tim Berners-Lee’s 5-star scale is a good reference here. The higher you can achieve on that scale, the more power of the Web you are using, the more you are facilitating reuse.

Are there differences regarding the use of LOD principles and technologies between developing countries and industrialised countries in your opinion? For example: does it make sense to start an Open Data Initiative in a developing country using Linked Open Data from the scratch?

All the issues with LOD I mentioned above apply and are even more strongly found in the developing world. I think we should take a step by step approach and start going from no data to some-star data in the very near term, lower the barriers one by one and start to building capacities in government and civil society but always with Web architecture principles in mind.
We will have to address the specificities of the developing world. For example, given that the LOD community is relying more and more on cloud-based options, on centralized data stores that require stable high-speed internet, how would one deploy a LOD solution in a country where clients (computers/mobile phones) have limited resources (disk, cpu) and where connectivity is unstable and with low-bandwidth? We’re participating in a worskhop to explore these issues.

This does not mean that LOD is completely ruled out from the beginning. As I pointed out before, there are cases on which it can be extremely useful and powerful and in those, we intend to accelerate adoption, likely piloting and building capacities as a first step.

Could you please tell us a few words about the Web Foundation?

The Web Foundation was launched by the inventor of the Web, Sir Tim Berners-Lee, in 2009 to address global challenges by connecting humanity and empowering individuals through an increasingly inclusive and powerful Web. More on the vision of the Web Foundation at:
http://www.webfoundation.org/vision/

Jose, many thanks for this interview. It seems that there is a quick progress in open data in developing countries as well as there are different requirements there to be taken into account in comparison to open data projects in Australia, the US or in Europe! Also the potential of Linked Open Data seems an interesting point for these countries!
We are looking forward to staying in touch with you on this in the future and wish you all the best for your future work in this area!

Step 1. Generate a thesaurus by using a linked data source like DBpedia

As recently reported SWC has developed a tool called SKOSsy which can be used to extract seed thesauri from DBpedia. In our example I will generate a knowledge model describing the domain of “digital photography“. This step took around 15 minutes.

Step 2. Load the thesaurus into PoolParty and improve it to your needs

After the seed thesaurus has been loaded into PoolParty Thesaurus Manager you have many possibilities to enhance the knowledge model further: Add more categories, synonyms, relations etc. In this example I use the seed-thesaurus without any further improvements. This step took approximately 2 minutes.

Step 3. Generate an automatic text extractor on top of your thesaurus

This step took a couple of seconds and ended up in having generated a fast and reliable text mining application on top of PoolParty Extractor, ready to be used to enrich your documents with data from the LOD cloud.

You can try it out here: PPX Live-Demo

To try the extractor on your own, please take a look at the image above which shows a proper configuration, you have to insert the following UUID in the form: d35d4ddb-adc3-4ea5-b027-deacac03e391

Since our example is all about ‘digital photography’, we recommend to use text samples (or some fragments) like these ones to test the quality of PPX based text analytics:

• Digital Camera Image Noise (Results as HTML, RDF/XML)
• Nikon D3S In-depth Review (Results as HTML, RDF/XML)
• Introduction to Shutter Speed in Digital Photography (Results as HTML, RDF/XML)
• Digital Camera Sensors (Results as HTML, RDF/XML)

Let us know what you think about this straight-forward approach and your opinion about the quality of the results. We believe that thesaurus based text mining is in many cases an alternative to some other approaches, especially if you want to to enrich your content with information from the upcoming web of data.

Of course we would be happy to generate other demos in the areas of your interest! Just get in contact with us by using our contact form.

Linked Open Data: The Essentials provides answers to the following key questions:

• What do the terms Open Data, Open Government Data and Linked Open Data actually mean, and what are the differences between them?
• What do I need to take into account in developing a LOD strategy?
• What does my organisation need to do technically in order to open up and publish its datasets?
• How can I make sure the data is accessible and digestible for others?
• How can I add value to my own data sets by consuming LOD from others?
• What can be learned existing best practices?
• What are the key potentials of sharing and consuming open datasets?

Read more about this publication and find out how to obtain a copy.

How to enter the SKOSsy-lottery:

• Enter a comment in this post. One comment per person. Describing which type of thesaurus you are interested in.
• All comments must be submitted before Jan 25, 2012.
• The winners will be selected at random.

Together with our PoolParty Suite, we are ready to present SKOSsy on our booth at SemTechBiz Berlin 2012 Exhibition area.  SKOSsy is a handsome tool, which generates SKOS based seed-thesauri in German or in English by extracting data from DBpedia. See our finger exercise on a thesaurus describing the world of Alan Turing – done with SKOSsy.

Let us know, which knowledge realm you are interested in and join the lottery now. Good luck, and see you in Berlin.