Difference between revisions of "Blue Hackathon iMarine Data Challenges"
(→Objectives) |
(→Background) |
||
| Line 5: | Line 5: | ||
Most information resources in the "Blue" domain were created without the exploitation by advanced search and discovery mechanisms in mind. They thus lack the semantic richness that would improve their visibility, usefulness, and quality. | Most information resources in the "Blue" domain were created without the exploitation by advanced search and discovery mechanisms in mind. They thus lack the semantic richness that would improve their visibility, usefulness, and quality. | ||
| − | One cost-effective opportunity to overcome this limitation may be the addition of | + | One cost-effective opportunity to overcome this limitation may be the addition of RDFa to existing datasets. This can be achieved by a mechanism that extracts concepts from a html-text, aligns these with concepts from a semantic KB, and returns the uri's that can be attached to the source, either off-line, as header metadata, or in-line. |
Proving that such a mechanisms can effectively enrich a 'flat' resource with interpretable rdf will present evidence for data owners in the "Blue" domain that they can add value to their resources with limited costs with the help of semantechnicians. | Proving that such a mechanisms can effectively enrich a 'flat' resource with interpretable rdf will present evidence for data owners in the "Blue" domain that they can add value to their resources with limited costs with the help of semantechnicians. | ||
Revision as of 12:33, 28 June 2013
Data Challenges
Challenge #1
Enrich HTML web content with RDF annotation, and enable annotation-based document discovery
Background
Most information resources in the "Blue" domain were created without the exploitation by advanced search and discovery mechanisms in mind. They thus lack the semantic richness that would improve their visibility, usefulness, and quality.
One cost-effective opportunity to overcome this limitation may be the addition of RDFa to existing datasets. This can be achieved by a mechanism that extracts concepts from a html-text, aligns these with concepts from a semantic KB, and returns the uri's that can be attached to the source, either off-line, as header metadata, or in-line.
Proving that such a mechanisms can effectively enrich a 'flat' resource with interpretable rdf will present evidence for data owners in the "Blue" domain that they can add value to their resources with limited costs with the help of semantechnicians.
Objectives
We ask the hackathon participants to find a technical solution to enrich the factsheets of the FIGIS portal with annotations in RDFa format. The annotation will consist at least the URIs of the entities referenced in the factsheet, and of set of relevant relations provided with the datasets.
We ask the hackathon to:
- GOAL: Provide an RDFa client to 1. extract concepts from fact-sheets, 2. identify uri's from several KB's, 3. create the RDF annotations, and 4. expose these RDF annotations.
- GOAL: Use the annotations produced at item one, as input to online search of factsheets (publication, GIS maps, images, statistical timeseries), to create enhanced discovery facility that complement the web page information content.
- GOAL: Retrieve a set of fact-sheets via online search services.
- GOAL: Write RDFa to these factsheets.
Challenges
TBD
Datasets
TBD
APIs
TBD
Challenge #2
Generate RDF dataset from GIS layer in geonetwork, and map the geographic entities with existing LOD datasets
Background
(Why this is relevant to blue-er world)
Objectives
- We ask the hackathon to find a technical solution to produce LOD dataset from a collection of GIS layers accessed via GeoNetwork web services. The entities of the dataset will have to be mapped with existing LOD datasets in the GIS domain.
- GOAL: given an online service that list a collection of GIS layer, a new LOD dataset is produced.
- GOAL: enrich the geographic entities in that dataset with more data gathered trough the mapping with existing LOD GIS datasets (e.g. geonames, geopolitical ontology, dbpedia, etc).
Challenges
TBD
Datasets
TBD
APIs
TBD
Challenge #3
Generate RDF dataset from DarwinCore sources and map to existing biodiversity LOD
Background
(Why this is relevant to blue-er world)
Objectives
- We ask the hackathon participants to produce a LOD dataset from a source of DarwinCore data (XML or a service), and map its entities with existing LOD datasets in the biodiversity domain.
- GOAL: access complementary information with taxonomic data through the mappings (e.g. species conservation status, capture statistics, distribution map, etc)
Challenges
TBD
Datasets
TBD
APIs
TBD
Challenge #4
Generate dynamic fact-sheets mashing up data from distributed LOD datasets
Background
(Why this is relevant to blue-er world)
Objectives
- We ask the hackathon to find a technical solution based on LOD data mashup, to compose domain-based sections of a factsheet, taking data from distributed LOD datasets. The domain of the sections can be: economics, taxonomic, fishing technique, statistics, publications etc.
- GOAL: a web service responding with a collection of data clustered by domain-section, and display the result in HTML format
Challenges
TBD
Datasets
TBD
APIs
TBD
Challenge #5
Search Results presentation exploitation.
Search results regarding marine data could be enriched in order to provide advanced experience to the user. Derived information could be injected into results regarding identification of special keywords (related to the query), with results retrieved by OpenSearch and other external(?) datasources. Also exploration of the results could be improved from simple browsing into information discovery, providing accumulated information, filtering, suggestions etc.
Background
(Why this is relevant to blue-er world)
Objectives
- We ask the hackathon participants to enrich the search results retrieved from iMarine Collections by identifying special keywords (related to the topic) with results retrieved from OpenSearch and other external(?) datasources.
- We ask the hackathon participants to explore the database by performing a number of predefined queries and keep statistics on them in order to enhance the existing browsing methods
Challenges
TBD
Datasets
APIs
Challenge #6
Processing and Visualization of data sets
Exploit geolocation of real-world data in order to calculate and visualize geographical information and trends (i.e. migration of species). Support interactive map search over multiple sources, combined and enriched results. Search results will be presented on a map with possible options of clustering, filtering etc. User could also interact with results, like clicking on a result or location would show related results, helpful things etc.
Background
(Why this is relevant to blue-er world)
Objectives
- Exploit the species occurrences data in order to calculate and visualize geographical trends (i.e. migration of species).
- Interactive Map Search. Search over data of multiple source, combine them and enrich results. Search results can be presented on a map.
- clustering, filtering
- trend identification
- interact with results, like clicking on a result or location would show related results, helpful things etc
Challenges
TBD
Datasets
APIs
Some Notes from FORTH about possible challenges
What is this fish? Shoot and learn
Title: What is this fish? Shoot and learn The user takes a shot of a fish using his mobile. The application uses the images of species (e.g. by exploiting ECOSCOPE’s images) and returns to the user the name of the fish and related information. Requirement: image similarity.
Gradual Query Expansion for Species (semantic pre-processing of keyword queries)
Title: Gradual Query Expansion for Species (semantic pre-processing of keyword queries) Challenge: tackle the problem of empty (or small) answers in search systems by designing and developing a component that allows gradual query expansion which exploits the availability of linked data. Input: A species name, a number of related sources of information Output: A series of queries <q1, q2, … , qk>, where a query is a set of words. The words in q_i is subset of the words in q_{i+1} , and so on. For instance q1 could contain the names of the species in different natural languages, q2 could include the scientific names, q3 could include sub/sup-species, q4 could include competitors, predators, etc. It could be deployed as a web app where the user enters his query, the app computes the expanded queries and could directly forward the control to a search engine (the expanded query is passed through the url).
Linked Data for Species
Title: Linked Data for Species Link the species described in the TLO-based warehouse (SPARQL endpoint) with related information in other sources of structured (e.g. DBPEDIA, ..) or unstructured information (e.g. Wikipedia, …) aiming at ….<<we need a specific objective here>>
Linked Data Browser
Title: Linked Data Browser Build a browser (textual/graphical) for the TLO-based repository (SPARQL endpoint). You should consider devices with small screens (one could develop a dedicated android client for this). Other SPARQL endpoints (or structured information accessed through HTTP) could also be considered. Challenge: tackle overloading
Datasets and APIs
Datasets
TLO based SPARQL endpoint
Data Graph
Description
The description of the MarineTLO can be found here:
http://wiki.i-marine.eu/index.php/Top_Level_Ontology
Exploitation Example
(How can be used within a challenge)
FAO FLOD
Description
The Fisheries Linked Open Data (FLOD) stems from a rising trend initiative known as Linked Open Data. It is dedicated to create a dense network of relationships among the entities of the Fishery domains, and to programmatically serve them to semantic and traditional application environments. It started with the objective to identify and interlink equivalent codes from different code lists in use by FIGIS, in order to consolidate the information referenced by each different code, and then expanded to include external data source such as NAFO, EU, and ICCAT. Currently the FLOD network includes entities and relationships from the the domains of Marine Species, Water Areas, Land Areas, Exclusive Economic Zones. It serves software applications in the domain of statistics, and GIS. The FLOD content is exposed via either SPARQL endpoints (suitable for semantic applications), or via JAVA API to be embedded in consumers' application code. The future work for FLOD includes the definition of guidelines to connect external interlinked data sources, to advance the current status of existing information systems by developing semantic services in response to their requirements, and to integrate FLOD in a corporate linked data network under discussion in FAO.
Exploitation Example
(How can be used within a challenge)
Ecoscope
Knowledge base on Exploited Marine Ecosystems
Opensearch description document
iMarine GeoNetwork
Description
(Short description of available data)
Exploitation Example
(How can be used within a challenge)
iMarine Biodiversity Data Service
Endpoint of the WS giving access to Biodiversity data coming from several providers ( OBIS, GBIF, CoL..)
Description
(Short description of available data)
Exploitation Example
</pre>
APIs
SPARQL Client
Any SPARQL client available on the Web
Description
(documentation)
Exploitation Example
(How can be used within a challenge) + (javadoc?)
GeoNetwork Client
Description
Wiki
A library to interact with GeoNetwork's REST Interface to publish/modify/delete and search for Metadata.The library is designed on top of geoserver-manager library, developed by GeoSolutions. Metadata objects managed by the library are compliant to standard specification ISO 19115:2003/19139.
Exploitation Example
Javadoc
SPD Client
Description
Wiki
The SPD Client can be used to access a Biodiversity data broker implemented in iMarine, the SPD service. More details about the architecture of the service are available at
https://gcube.wiki.gcube-system.org/gcube/index.php/Biodiversity_Access
Exploitation Example
The client can be used for example to query the OBIS data source and return the taxonomic information related to shark
ScopeProvider.instance.set("/d4science.research-infrastructures.eu/gCubeApps");
Manager manager = manager().withTimeout(3, TimeUnit.MINUTES).build();
Stream<ResultElement> taxa = manager.search("SEARCH BY CN 'shark' RESOLVE WITH OBIS EXPAND IN OBIS RETURN Taxon");
while (taxa.hasNext()){
TaxonomyItem taxon = (TaxonomyItem)taxa.next();
System.out.println(taxon.getAuthor()+" "+taxon.getRank()+" "+taxon.getScientificName());
while ((taxon=taxon.getParent())!=null)
System.out.println(taxon.getScientificName()+" -- "+taxon.getRank());
}
gCUbe Search client
Description
Wiki
Exploitation Example
Javadoc
Artifacts
The software distributed by iMarine ( gCube ) is available trough Maven repositories. The following setting.xml configuration file should be set up:
<settings xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/settings-1.0.0.xsd"> <profiles> <profile> <id>gcube</id> <repositories> <repository> <id>gcube-releases</id> <name>gCube Releases</name> <url>http://maven.research-infrastructures.eu/nexus/content/repositories/gcube-releases</url> <releases> <enabled>true</enabled> </releases> <snapshots> <enabled>false</enabled> </snapshots> </repository> <repository> <id>gcube-externals</id> <name>gCube Externals</name> <url>http://maven.research-infrastructures.eu/nexus/content/repositories/gcube-externals</url> <snapshots> <enabled>false</enabled> </snapshots> <releases> <enabled>true</enabled> </releases> </repository> </repositories> <pluginRepositories> <pluginRepository> <id>gcube-releases</id> <name>gCube Releases</name> <url>http://maven.research-infrastructures.eu/nexus/content/repositories/gcube-releases</url> <releases> <enabled>true</enabled> </releases> <snapshots> <enabled>false</enabled> </snapshots> </pluginRepository> <pluginRepository> <id>gcube-externals</id> <name>gCube Externals</name> <url>http://maven.research-infrastructures.eu/nexus/content/repositories/gcube-externals</url> <snapshots> <enabled>false</enabled> </snapshots> <releases> <enabled>true</enabled> </releases> </pluginRepository> </pluginRepositories> </profile> </profiles> <activeProfiles> <activeProfile>gcube</activeProfile> </activeProfiles> </settings>
or the same settings included in your pom file. The maven coordinates of the components to use for the challenges are documented in the related wikis.
