Archive

Posts Tagged ‘Linked Open Data’

Pourquoi utiliser le Web de données?

Il y a quelque jours j’ai eu le plaisir, et la chance, de participer à la série de webinaires organisés par l’AIMS. L’objectif que je m’étais fixé pour ma présentation (en Français) intitulée “Clarifier le sens de vos données publiques avec le Web de données” était de démontrer l’avantage de l’utilisation du Web de données du point de vue du fournisseur de données, en passant par le consommateur. Faire une présentation sans aucun retour de la part de l’auditoire était une expérience intéressante que je renouvèlerait volontiers si une nouvelle occasion se présente. Surtout si c’est Imma et Christophe qui sont aux commandes! grâce à eux tout était parfaitement organisé et le wébinaire s’est déroulé sans problème 🙂

Si vous voulez voir si cette présentation atteint son but, les diapositives sont disponible sur Slideshare:

Une autre copie de cette présentation est disponible sur le compte SlideShare de l’AIMS.

Advertisements

5-stars Linked Open Data pays more than Open Data

Let’s assume you are the owner of a CSV file with some valuable data. You derive some revenue from it by selling it to consumers that do traditional data integration. They take your file and import it into their own data storage solution (for instance, a relational database) and deploy applications on top of this data store.

Traditional data integration

Data integration is not easy and you’ve been told that Linked Open Data facilitates it so you want to publish your data as 5-star Linked Data. The problem is that the first star speaks about “Open license” (follow this link for an extensive description of the 5-star scheme) and that sounds orthogonal to the idea of making money with selling the data :-/

If you publish your CSV as-is, under an open license, you get 3-stars but don’t make money out of serving it. Trying to get 4 or 5 stars means more effort from you as a data publisher and will cost you some money, still without earning you back any…

Well, let’s look at this 4th star again. Going from 3 stars to 4 means publishing descriptions of the entities in the Web. All your data items get a Web page on their own with the structured data associated to them. For instance, if your dataset contains a list of cities with their associated population every of this city as its own URI with the population indicated in it. From that point, you get the 5th star by linking these pages to other pages published as Linked Open Data.

Roughly speaking, your CSV file is turned into a Web site and this is how you can make money out of it. Like for any website, visitors can look at individual pages and do whatever they want with them. They can not however dump the entire web site into their machine. Those interested in getting all the data can still buy it from you, either as a CSV or RDF dump.

Users of your data have the choice between two data usage process: use parts of the data through the Linked Open Data access or buy it all, and integrate it. They are free to choose the best solution for them depending on their needs and resources.

Using Linked Open Data

Some added side bonuses of going 5-star instead of sticking at 3:

  • Because part of the data is open for free, you can expect to get more users screening it and reporting back errors;
  • Other data publishers can easilly link their data set with yours by re-using the URIs of the data items. This increases the value of the data;
  • In its RDF format, it is possible to  add some links within the data set. Thereby doing part of the data integration work on the behalf of the data consumers – who will be grateful for it!
  • Users can deploy a variety of RDF-enabled tools to consume your data in various ways;

Sounds good, doesn’t it? So, why not publishing all your 3-star data as 5-star right away? 😉

Downscaling Entity Registries for Poorly-Connected Environments

VeriSign logo

VeriSign logo (Photo credit: Wikipedia)

Emerging online applications based on the Web of Objects or Linked Open Data typically assume that connectivity to data repositories and entity resolution services are always available. This may not be a valid assumption in many cases. Indeed, there are about 4.5 billion people in the world who have no or limited Internet access. Many data-driven applications may have a critical impact on the life of those people, but are inaccessible to those populations due to the architecture of today’s data registries.

Examples of data registries include the domain name registries. These are databases containing registered Internet domain names. They are necessary for all Web users wishing to visit a website knowing its URL (e.g. https://semweb4u.wordpress.com) rather than its IP address (e.g. http://76.74.254.120). Another example of data registry is the Digital Object Architecture (DOA) which assigns unique identifiers to digital objects (e.g. scientific publications).

Registries are critical components of today’s Internet architecture. They are widely used in every-day Web activities but their usage is severely impaired in poorly connected or ad-hoc environments. In this context, centralized data management – as typically used by current data registries – is of limited practicability, if only possible in the first place. There is a need for hybrid models mixing decentralized and hierarchical infrastructures to support data-driven application in environments with limited Internet connectivity.

Philippe Cudré-Mauroux and myself, received a $200,000 research grant from VeriSign Inc. (PDF version) to investigate such novel approaches for data registries. During this 12 months project, we will develop decentralized solutions to the problems of entity publication, search, de-duplication, storage and caching. A running prototype will be tested on the XO laptop, a laptop used by young learners in developing countries – most often in a mesh context with limited Internet connectivity.

Please don’t hesitate to contact us to ask for information about this project, we’d be happy to talk more about our plans 🙂

Exposing API data as Linked Data

The Institute of Development Studies (IDS) is a UK based institute specialised in development research, teaching and communications. As part of their activities, they provide an API to query their knowledge services data set compromising more than 32k abstracts or summaries of development research documents related to 8k development organisations, almost 30 themes and 225 countries and territories.

A month ago, Victor de Boer and myself got a grant from IDS to investigate exposing their data as RDF and building some client applications making use of the enriched data. We aimed at using the API as it is and create 5-star Linked Data by linking the created resources to other resources on the Web. The outcome is the IDSWrapper which is now freely accessible, both as HTML and as RDF. Although this is still work in progress, this wrapper already shows some advantages provided by publishing the data as Linked Data.

Enriched data through linkage

When you query for a document, the API indicates you the language in which this document is wrote. For instance, “English”. The wrapper replaces this information by a reference to the matching resource in Lexvo. The property “language” is also replaced by the equivalent property as defined in Dublin Core, commonly used to denote the language a given document is wrote in. For the data consumer, Lexvo provides alternate spelling of the language name in different languages. Instead of just knowing that the language is named “English”, the data consumer, after deferencing the data from Lexvo will know that this language is also known as “Anglais” in French or “Engelsk” in Danish.

Part of the description of a document

Links can also be established with other resources to enrich the results provided. For instance, the information provided by IDS about the countries is enriched with a link to their equivalent in Geonames. That provides localised names for the countries as well as geographical coordinates.

Part of the description of the resource "Gambia"

Similarly, the description of themes is linked with their equivalent in DBpedia to benefit from the structured information extracted from their Wikipedia page. Thanks to that link, the data consumer gets access to some extra information such as pointers to related documents.

Part of the description of the theme "Food security"

Besides, the resources exposed are also internally linked. The API provides an identifier for the region a given document is related to. In the wrapper, this identifier is turned into the URI corresponding to the relevant resource.

Example of internal link in the description of a document

Integration on the data publisher side

All of these links are established by the wrapper, using either SPARQL requests (for DBpedia) or calls to data API (for Lexvo and Geonames). This is something any client application could do, obviously, but one advantage of publishing Linked Data is that part of the data integration work is done server side, by the person who has the most information about what his data is about. A data consumer just as to use the links already there instead of having to figure out a way to establish them himself.

A single data model

Another advantage for a data consumer is that all the data published by the wrapper, as well as all the connected data sets, are published in RDF. That is one single data model to consume. A simple HTTP GET asking for RDF content returns structured data for the content exposed by the wrapper, and the data DBpedia, Lexvo and Geonames. There is no need to worry about different data formats returned by different APIs.

Next steps

We are implementing more linking services and also working on making the code more generic. Our goal, which is only partially fullfiled now, is to have a generic tool that only requires an ontology for the data set to expose it as Linked Data. The code is freely available on GitHub, watch us to stay up to date with the evolution of the project 😉

Take home messages from ePSIplatform Conference

Open Data stickers

Open Data

On March 16, 2012 the European Public Sector Information Platform organised the ePSIplatform Conference 2012 on the theme “Taking re-use to the next level!”. A very well organised and interesting event, also a good opportunity to meet new persons and put a face on the names seen on the mails and during teleconferences 🙂

The program was intense: 3 plenary sessions, 12 break-out sessions and project presentations during the lunch break. That was a lot to talk about and a lot to listen to. I left Rotterdam with a number of take out messages and food for thought. What follows is a mix of my own opinions and things said by some of the many participants/speakers of the event.

We need to think more about data re-use

It’s a fact: Open Data has reached momentum and more and more data portals are being created. DataCatalogs currently lists 216 sources for Open Data. There could be something around a million of Open Data data sets now available, but how many applications? Maybe around 100k, at most. Furthermore, most on these applications do not really address “real problems” (e.g. help deciders to make educated choices by providing them with the right data at the right time, or optimise food distribution processes). Even if the definition of a “real problem” is open to discussion, there is surely something to think about.

This low number of applications could be explained by a lack of problems to tackle as well as it can be explained by a lack of motivated developers. The ePSI platform has just started a survey on story sharing. Reading about the (positive) experience of others is likely to trigger some vocations in the readers and get more developers on board. The upcoming W3C event about using Open Data will also be a good place to share such stories and spot the things to do next to foster an ecosystem of data and developers.

Open Data should be interactive

We have Open Data and we have Open Data consumers that happily take the data, process it and eventually re-publish it. Fine but we do poor when it comes to capture the added meta data from these users. If one of them spot an error in an open data set, or if missing data is identified, there is hardly any way to communicate this information to the data publisher. Most, if not all, data portals are “read only” and the eventual feedback they receive may not find a matching processing pipeline. Open source software solved this issue by using open bug trackers that allows for reporting bugs/feature requests and facilitate dispatching the issues to persons in charge of implementing them. Using such bug trackers to keep the data users in the loop sounds like a good plan. This is something we started to look at, in a slightly different way, for the projects CEDA_R and Data2Semantics. One of the use case of these projects is the Dutch historical census data (from 1795 onwards) that has to be harmonized and debugged (there was a lot of manual process involved to convert the paper reports in digital form). Only historians can take care of this, and they need to inform the data publisher about their finding – preferably using something even easier that the average bug tracker.

Open (messy) Data is a valuable business

Economical issues are common when speaking about Open Data. They could even be seen as the main obstacle to it. The other obstacles, technical, legal and societal/political being easier to address. So the trick is to convince data owners that, yes, they will loose the money they currently get in access fee but they will get more out of the Open Data, in an indirect way through businesses created. In fact, there is no market for the Open Data itself. Instead, this Open Data has to be seen as part of the global Data market of which DataPublica and OpenCorporates are two examples. In this market, curating and integrating data is a service clients can be charged for. Data companies transform the data into information and put a price tag on the process.  For this matter, having to publish an integrated data set as Open Data because it include pieces of an other Open Data set licensed with a GPL-like license will brake the process. Open Data is easier to consume when license under more BSD-like licenses.

If there is a market for messy open data,  one can wonder whether Linked Data is going against businesses or helping them. Linked Data allows for doing data integration at the publication level and Open Data published exposed using these principles is richer and easier to consume.  This means less work for the consumer, which may spare himself the cost of hiring someone to integrate the data. But Linked Data facilitates the job of data companies too. These could invest the time saved into the development of visualisation tools, for instance. So in the end, it may not be such a bad idea to continue promoting Linked Data 😉

Open Data initiatives need to become more consistent

Besides the definition given on OpenDefinition, and the 5-star scheme of Tim Berners Lee for Linked Data, there is not much out there to tell people what is Open Data and how to publish it. Data portals can be created from scratch or use CKAN and may expose the meta data about the data sets it contains in different ways (e.g. using DCAT or something else). The data itself can be published within a large spectrum of formats ranging from XLS sheets to PDFs to RDF. Besides this, data portals can be created at the scale of the city, a region, an entire country or an entity such as the EU. These different scales are related to each other and can be seen as a result from a lack of coordination. Directories are important as a way to know what data is out there, and also what data is missing. If everyone take initiatives at different scales, the outcome of this indexing process will be fuzzy and the outcome quite confusing for data users looking for open data sets. On the other hand, self-organisation is often the best solution to build and maintain complex systems (c.f. “Making Things Work” from Y. Bar-Yam). So maybe things are good as they are but we should still avoid ending up with too many data portals  partially overlapping and incompatible with each other.

As far as the data is concerned, PDF, XLS, CSV, TSV, … are all different ways to create data silos that just provide a single view over the data – even a non machine readable one in the case of many PDFs. RDF is here to improve consistency across data sets with a unique, graph based, data model. This data model facilitates sharing data across data sets. It is not the only solution to do that, the data set publishing language (DSPL) from Google being an other one, but it is the only one based on W3C standards. This guarantees the openness of the data format and a constant support, just as for the standards that make the Web (HTML, HTTP, CSS, …).

Don’t underestimate the “wow” effect

During one of the break-out sessions, I was intrigued hearing one of the panel speaker saying he would like to see more DSPL around than RDF. After some (pacific) discussion, we agreed on the following points: RDF is more expressive than DSPL, DSPL comes with an easy to use suite of plug&play tools to play with the data. It seems that if you want to re-use Open Data to do some plots, eventually for some data journalism use-cases, you are better off using DSPL. It is simpler and through the data explorer allows anyone to build graphs in a few clicks. Users prefer having button and sliders to play with simpler data rather than knowing that they have in their hands the most powerful knowledge representation scheme and that they could do anything with it – but finally do nothing with it because of the induced high learning curve. I’m all in favour of Open Data and I try to motive people, and myself sometime, to use Linked Data to publish data sets. Still, I think we have a major issue there: our data model is better but we do not compete yet on the usability side of the story.

An other manifestation of the “wow” effect: the most impressive visualisation show at the event was a part of the video documentaries “The Netherlands from above”, and their matching interactive data explorers.  This is a very nicely done job but the interesting bit is that not only the data was not linked, it was also not open! However, even at an event about re-use of Open Data, nobody seemed to care much. The data was acquired for free from different providers, with some difficulties for some, had to be curated and transcoded, and could not be shared. But the movies are very nice, and the sliders on the interactive pages fun to play with…

We must not rest on our laurels

Finally, and that was also the final message of the event, we should not rest on our laurels. Open data is well received. Many are going into the “Open unless” way of thinking but some others make an Open Data portal just because it is trendy, and trash it after some months. We need to continue explaining to data owners why they should open their data and explain why Linked Data is a good technical solution to implement. Then, we need to find more active users for the data because, in the end, if the data is used, nobody will even dare shutting down the portal serving it. Having these active users  may be our only guarantee that data published as Open Data will remain as such for the years to come.

Updates about SemanticXO

With the last post about SemanticXO dating back from April, it’s time for an update, isn’t it? 😉

A lot of things happened since April. First, a paper about the project was accepted for presentation at the First International Conference on e-Technologies and Networks for Development (ICeND2011). Then, I spoke about the project during the symposium of the Network Institute as well as during the SugarCamp #2. Lastly, a first release of a triple-store powered Journal is now available for testing.

Publication

The paper entitled “SemanticXO : connecting the XO with the World’s largest information network ” is available from Mendeley. It explains what the goal of the project is and then report on some performance assessement and a first test activity. Most of the information contained has actually been blogged before on this blog (c.f. there and there) but if you want a global overview of the project, this paper is still worth a read. The conference in itself was very nice and I did some networking. I came back with a lot of business card and the hope of keeping in touch with the people I met there. The slides from the presentation are available from SlideShare

Presentations

The Network Institute of Amsterdam organised on May 10 a one-day symposium to strengthen the ties between its members and to stimulate further collaboration. This institute is a long-term collaboration between groups from the Department of Computer Science, the Department of Mathematics, the Faculty of Social Sciences and the Faculty of Economics and Business Administration. I presented a poster about SemanticXO and an abstract went into the proceedings of the event.

More recently, I spent the 10 and the 11 of September at Paris for the Sugar Camp #2 organised by OLPC France. Bastien managed me a bit of time on Sunday afternoon to re-do the presentation from ICeND2011 (thanks again for that!) and get some feedback. This was a very well organised event held at a cool location (“La cité des sciences“), it was also the first time I met so many other people working on Sugar and I could finally put some faces on the name I saw so many time on the mailing lists and on the GIT logs 🙂

First SemanticXO prototype

The project developement effort is split in 3 parts: a common layer hidding the complexity of SPARQL, a new implementation of the journal datastore and the coding of diverse activities making use of the new semantic capabilities. All three are going more or less in parallel, at different speed, as, for instance, the work on activities direct what the common layer will contain. I’ve focused my efforts on the journal datastore to get something ready to test. It’s a very first prototype that has been coded starting with the genuine datastore 0.92 and replacing the part in charge of the metadata. The code taking care of the files remains the same. This new datastore is available from Gitorious but because installing the triple store and replacing the journal is a tricky manual process, I bundled all of that 😉

Installation

The installation bundle consists of two files, a “semanticxo.tgz” and a script “patch-my-xo.sh“. To install SemanticXO, you need to download the two and put them in the same location somewhere on your machine and then type (as root):

sh ./patch-my-xo.sh setup

This will install a triple store, add it to the daemons to start at boot time and replace the default journal by one using the triple store. Be careful to have backups if needed as this will remove all the content previously stored in the journal! Once the script has been executed, reboot the machine to start using the new software.

The bundle has been tested on an XO-1 running the software release 11.2.0 but it should work on any software release on both the XO-1 and XO-1.5. This bundle won’t work on the 1.75 has it contains a binary (the triple store) not compiled for ARM.

What now?

Now that you have the thing installed, open the browser and go to “http://127.0.0.1:8080”. You will see the web interface of the triple store which allows you to make some SPARQL queries and see which named graphs are stored. If you are not fluent in SPARQL, the named graph interface is the most interesting part to play with. Every entry in the journal gets its own named graph, after having populated the journal with some entries you will see this list of named graphs growing. Click on one of them and the content of the journal entry will be displayed. Note that this web interface is also accessible from any other machine on the same network as the XO. This yields new opportunities in terms of backup and information gathering: a teacher can query the journal of any XO directly from a school server, or an other XO.

Removing

The patch script comes with an uninstall function if you want to revert the XO to its original setup. To use it, simply type (as root):

sh ./patch-my-xo.sh remove

and then reboot the machine.

CKAN->network exporter for the LOD Cloud

The LOD cloud as rendered by Gephi

One year ago, we posted on the LarkC blog a first network model of the LOD cloud. Network analysis software can highlight some aspects of the cloud that are not directly visible otherwise. In particular, the presence of dense sub-groups and several hubs – whereas in the classical picture, DBPedia is easily perceived as being the only hub.

Computing network measures such as centralities, clustering coefficient or the average path length can reveal much more about the content of a graph and the interplay of its nodes. As shown since that blog post, these information can be used to appreciate the evolution of the Web of Data and devise actions to improve it (see the WoD analysis page for more information about our research on this topic). Unfortunately, the picture provided by Richard and Anja on lod-cloud.net can not be fitted directly into a network analysis software which expects a .net or CSVs files instead. Fortunately, thanks to the very nice API of CKAN.net it is easy to write a script generating such files. We made such a script and thought it would be a good idea to share it 🙂

The script is hosted on GitHub. It produces a “.net” file according to the format of Pajek and two CSV files, one for the nodes and one for the edges. These CSV can then easily be imported into Gephi, for instance, or any other software of your choice. We also made a dump of the cloud as of today and packaged the resulting files.

Have fun analysing the graph and let us know if you find something interesting 😉

 

%d bloggers like this: