Logo TIB AV-Portal Logo TIB AV-Portal

Integrating the spatial web with linked open data using GeoDCAT-AP

Video in TIB AV-Portal: Integrating the spatial web with linked open data using GeoDCAT-AP

Formal Metadata

Integrating the spatial web with linked open data using GeoDCAT-AP
Title of Series
Part Number
Number of Parts
CC Attribution 3.0 Germany:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Release Date
Production Place

Content Metadata

Subject Area
GeoDCAT-AP is an extension of DCAT-AP for describing geospatial datasets, dataset series, and services. It provides an RDF syntax binding for the union of metadata elements defined in the core profile of ISO 19115:2003 and those defined in the framework of the INSPIRE Directive. Its basic use case is to make spatial datasets, data series, and services searchable on general data portals, thereby making geospatial information better searchable across borders and sectors. This can be achieved by the exchange of descriptions of datasets among data portals.
Keywords GeoCat
presentation Blocks projects series simulation system call Arm open subsets metadata van variance training
point man web Slides standards flow maximal fractals open subsets Emulation Testbed domain web Computer animation rates case organization testing fuzzing web
presentation Actions time Linked Open Data ones sets open subsets iked programs web specific Densities profiles ons core spaces Dean-Zahl standards The list bits lines open subsets Computer animation case search engine form
regional maximal sets open subsets programs metadata fields number Pointer service goodness hypermedia ons core Rolling model extent Cats Sum descriptions domain Beats man standards distribution mapping analysis coma basis catalog bin information category Computer animation record
domain Linked Open Data Transformers civil real schemes sets open subsets catalog programs single category Computer animation single objects model
standards email formating Transformers experts The list bits catalog open subsets clients functions particle Computer animation website model Conversation protocols protocols record
choice area files mapping information Mass open subsets applications Location specific service inclusion Computer animation different website Conversation protocols
web pages Identify link Open Source open subsets metadata service goodness configuration string structure share exception addition standards formating cellular bits catalog Types Computer animation software search engine Query Results record
Identify link Query load box level catalog Conversation Cats metadata geometric
web building implementation formating integrators open subsets catalog limitations energy number Indexable means Computer animation software share addresses Results
web pages choice noise presentation information mapping link code time bits loss Computer animation case model Results
Computer animation
Cherry while you so in
a search session this block of calls and you can actually you will not stop linking the data and how do readers especially on him so welcome everybody and this this is actually 1 in a series of training we had we had 1 of the 2 presentations just before and it is 1 it was about big data and the role of metadata and but I'll come back to that later and some of this work has been done in the scope of the article where project which is in FP 7 European at 7 projects relating to citizen science and so the main thing that
I want to talk about is is is a recent test that they can overcome this set up here in Nova Mr. deduction the organization and Lin now virtually could be here today as he made this this data to indicate how disconnected we assess as spatial community are from the web and so here we have our heart rate uh OGC standards which work investigating how the mind but on the the other side of the wall doesn't really know what what we're doing here but do interest this all these terabytes of data items using 1 of the guys in doing so so we get a lot of in in in Dutch parliament there were questions from the Open Data can go to the to the Ministry of of infrastructure like like and we access your data so so did you would get went into the parliament and the case of them had to go up with a good answer well we have these OGC standards which ISO standard so that good but it was interesting discussion about heard and with
this slide here because this is recent work from from the w Tracy data publishing on the Web for the Working Group is you would have it we as communities would have followed these 30 is there would be no problem at all but there's a lot of them is is and home to read and at some point is is just as an person isn't in
it and so so um the to see 1 of very 1st go you see for for a conservative for communities actually which each of them uh density have its own data cloud so we have a spatial ones have our data cloud with our data specifications of vocabularies there's another 1 which is the the spatial lines in the spatial of this search engines in their own ontologies schools along and they have their own time and there unforgettable signal with the Google Knowledge Graph and then the being of and then there's the Open Data that and so that is 1 the Linked Data community they use this avoided ontology which is also a WGC ontology to to describe datasets and they use the Nupedia OpenStreetMap as their core vocabulary and you noticed that picture about the the the linked web with all the uh the green and pink of data sets and typically 1 here it has to be PDI in the center and then there is this open data thing what is that in Guinea get is an RDF standard so it should be here but um you see that within Europe that place Eurovoc here as the core ontology you notice that a lot of do get and how the get is its own national countries implementing national profiles on top of the the again they place um and legal vocabulary on here is the core of their now which is different than this 1 so this is not the same thing and then we have lots ISO-code lists as the core of ontologies so we have 4 communities which are kind of disconnected and at the end as a died provided the only thing I can give to you is if you identify each of these use cases as your audience then try to service all of them and to do that
so these are individually gets a little bit EU relevant that groups there's probably a lot more but these are 1 is something we work with so this is a big 1 in the eyes of program from the European Union Arena did a lot of work related to mapping ISO to together this is 1 which is mentioned in the previous presentation if initiative in the Netherlands the lessons of and that's important 1 that's the global OGC and w 3 c having a working group now which coordinates all that the work done in in setting up a separate assisted to to combine them so this is the
again is actually quite simple as that it's similar in the in the basis as I said this is no size 0 1 I want to know is just to be sure I to explain who does not know it 1 2 3 so I the 1 I want to is is smallest indeed Media metadata standard that we use in the spatial domain to describe data sets of features of services and it is that I don't have a picture of it it's a similar model as this in the basis but it has a lot of extensions which are totally different and this 1 and then so this is the cat is kind of the DP core ontology used globally then was to get out of which is in the European Union and extension of this one's a little couple of extra fields uh 40 European Open Data Portal and then this 1 was is use as a program GOOD et al which has added another set of properties and To facilitate also the mapping between inspire metadata and gets metadata so there's not 1 D against those there's of extensions but what what is good
good to know where is the center is a get catalog defined so you have a catalog which contains datasets and each of those datasets and a number of distributions united so that will be datasets having analysis and you get a catalog records which hold hold the description of the dataset the
then so what what was developed in the eyes of program is transformation from ISO 1 9 1 2 9 2 the the goal was to to make our either make the data available in the open data catalogs which usually are based on together so that it is nice picture of my daughters who represents something else what is that
if you do such a transformation from 1 model to another model you arrive and challenges and 1 of those challenges here is is here is the guns arbitration to go the negotiation is typically something from from the Linked Data Domain uh if you ask for a certain civilization of a certain documents and you don't specify which cannot you wanted it so what a couple of recommendations did was still too hardly certain scheme certain the performance of those 2 to bring to overcome that that's the capability is a typical practical solution that that that arise when you do work like this when I when I showed this to to some real RDF people eagerly wanting in doing here this is where because for them that they're used to have in a single RDF documents to say OK this object that i'm describing is as well it you get a dataset documents and avoid documents that this that and the more data sets and then just list the properties which are valid for each of these so so in the idea of did this to solve it but In the other domains we we always tend to transform what things to another before we can use it and then this so so
just a transformation is not all we also need protocols in uh in the eyes of the users of the world would we use it to do and see a summary so typical request from the urge to well that from an OGC kind of client OK they you provide do you get as an output format for C is only because used dealing with C on you but if you look at that from the other side was that was going to use that they that they don't know what the is w they only know get coming from spiral or from some but but other sites so it is a bit of a strange question that operate you last we build so so you know what you have this capability now to do experts any ISO records as the on easy you get is more and more useful to have this model as a particle but then if you look at a typical open data catalogs most of them don't have sparkled endpoints right now they do applied to get model but they don't have a SPARQL endpoint what they had discussed maybe Ice Age open data catalog out there has its own API which he won a connective each of these open data catalogs you have to learn a new API instead using the OpenID community is quite new and that they they're still struggling with with setting up the standards I know that they're
going to the conversion challenges there's a lot I'm not going to jail with this is follow the mailing list is very good for example there's
1 there is a need to get as many examples and you're on axis URL this specification doesn't really tell you what the difference between 3 and I usually say if it's WMS WFS festivities service on it admitted nexus euro and dollar dural if if it's a file and you download it then it's it's a done deal I so doesn't really require you to put uh a protocol or something you can and I would recommend you to put a protocol and so if you have a WFS URL mentioned that it's a WFS Europe so so you can make a good choice in the conversion and then
W map so it the European Open Data Portal as a lot of WMS website service endpoints but is doubly must Open Data I don't you can do it you can view of feature info everywhere received there's area but it's that I personally I would not include w mass as Open Data in any Open Data Portal what this is about the
axis here as in the in the OTC will vary we would tend to put uh services service and it just the plain service your but if decide if this data gets harvested by an open data portal and so we click on the link gets this 1 No . 4 exception you forgot to put services capabilities there so when you write your metadata mind is type of things this this metadata may be harvested by portal which is not aware about the standards and you could add to the question is cells or maybe people get feature requests for each of the formats that are available in the and then the next 1 is the of the approach Approach and I'm going to talk about it separately this is 1 this is a typical as the model and and if you put structure layer on top of that this is the work that that Clemens and I have been doing it in in the scope of this you flagged as bad we said OK you don't know how wfsc will give you a proxy W of as HTML or as seasonal adjacent so so any node with this L the prox software and you the enable people to to access NEW fast using their standards being the Eisenhower or for search engines and additional pages I'm gonna write it in the next session which is tomorrow I'm going to now go a bit more into this approach but it is as a result each WFS records in European which is linked to a catalog records will show a nation of age which can be crawled by search engine and and there's also the addition of the 2 adjacent and you UML options export options this is daily practice over is open source you can hear you integer try a darker instances and try it out tonight in your hotel room
identifiers so so linking to other things it's important to know whether it makes it makes it makes it it allows you and it allows a lot of people can find you and allows us to do interesting SPARQL queries like Give me all people of them where where and in Italy to welcome you to use your eyes goodness values keywords don't just put strings that put your eyes it makes the converted to do get easier and better the value of the different increases
in ISO 1 9 1 2 9 you can have geometry as an identifier so instead of putting a box for the datasets is say OK this is data about Germany are by means and and then use the the identifier of mentions from the DVD or from uh 1 of the official good you European and the cat
requires that approach so that that makes the conversion easier so we did experiment already a year ago to the national catalog you can be converted easily requisite began a loaded level individuals so and query in the metadata using spot is great works fine you can do with what we could find a dataset but then it stops because sparkle there's no but w fast advisor WVS link but doesn't go into the data so recently
energy of where does that with that the LTTE proxy approach could go a step further this this 1 this is these results are actually published at the beginning of this week love love love is is a harvester and which have this data and this 1 was able to harvest all the WFS data out why at the old and the proxy approach as RDF so you an amazing number of triples there because that's the full at the harvest all where all triples on the Web their harvests into the index and now they're adding are WFS data from the Dutch has the and you can query is a lot longer amount as 1 so the full Dutch as the I will be available advise that's amazing addresses everything we have done that doesn't mean we really make sense for like uh uh coverage datasets but but from things like addresses and the buildings and then makes it makes a lot of sense to have this and so
do you get and you network so we have this this implementations and the 2012 uh so recently added the atmosphere so what's next is a lot of things so see canisters to be 1 of the big open data catalogs so we want to improve the integration with that now we converted but we also want to convert and again to ISO so you are also able to ingest that will the and all the formats like limitation of the intuitive so get a red both
for coffee and what we see in these mappings you you have to make a lot of choices because those models don't don't fully match the is either in the properties or in the code his values that you use and it's always important to find ways a clear use cases what is the use case that I'm doing this transformation for because if it evident you will make other choices if you if you use this is different and then we sort as for communities there's not a single approach due to the satisfied each of them so going back to that you probably have to set up for endpoints due to their or maybe 5 or 6 if you identify more communities to to facilitate each of them convergence always the result in data loss is like that is bad for assessment but for discovering fine use a little bit people in the end will click to the original resource and find the information there and is not simple and Click on the link at home this is again an issue which goes on for 4 or 5 pages it's it's not to bring these wells together but having these conferences like this I I think we progress and so does anybody have
any questions for all the all your local questions to a question for yourself that that that that Mary and he has focused on how to OK because also tomorrow it as a matter of a surgeon OK well what time is off noise presentation and try to a telephone the Paul