=Paper=
{{Paper
|id=Vol-1501/Diversity2015-paper_1
|storemode=property
|title=From Data Portal to Knowledge Portal: Leveraging Semantic Technologies to Support Interdisciplinary Studies
|pdfUrl=https://ceur-ws.org/Vol-1501/Diversity2015-paper_1.pdf
|volume=Vol-1501
|dblpUrl=https://dblp.org/rec/conf/semweb/MaWEZCWZF15
}}
==From Data Portal to Knowledge Portal: Leveraging Semantic Technologies to Support Interdisciplinary Studies==
https://ceur-ws.org/Vol-1501/Diversity2015-paper_1.pdf
From data portal to knowledge portal: Leveraging
semantic technologies to support interdisciplinary studies
Xiaogang Ma, Patrick West, John Erickson, Stephan Zednik, Yu Chen, Han Wang,
Hao Zhong, Peter Fox
Tetherless World Constellation, Rensselaer Polytechnic Institute, Troy, NY, USA
Email: {max7; westp; erickj4; zednis2; cheny18; wangh17;
zhongh3; foxp}@rpi.edu
Abstract. Scientific research practices regularly adopt new technologies and
platforms in an effort to increase information timeliness, sharing and discovera-
bility. There are many initiatives related to open data, open code, open access,
open collections, composing the topic of Open Science in academia. Being open
has two levels of meanings. The first is to make the data, code, sample collec-
tions and publications, etc. freely accessible online. The other is the annotation
and connection between those resources to establish the provenance information
for reproducible scientific research. In this paper we present our work on a web
portal for the Deep Carbon Observatory (DCO) community [1]. The DCO is a
10-year (2009-2019) initiative to intensify global attention and scientific effort
in the burgeoning field of deep carbon science. Inspired by guiding questions
such as “how much carbon does Earth contain?”, “where is it?” and “what can
deep carbon tell us about origins?” more than 1000 scientists across the world
are actively participating in the DCO community. The DCO web portal is a re-
search collaboration website developed to keep track of all researchers, organi-
zations, instruments, field sites, and research outputs related to the DCO com-
munity. We intend for the DCO web portal to be a knowledge portal - adopting
state-of-the-art semantic technologies to support various stages of the scientific
process within and beyond the DCO community.
Keywords: Semantic Web; eScience; Knowledge Portal; Ontologies; Data
Stewardship
1 A model of the science network
The context of our work is the Semantic Web, which is defined as an extension to the
current Web by adding machine readable meanings and context to information on the
Web. In this way, the Web is being transformed from a Web of Documents to a Web
of Data [2]. Ontologies are an important way of capturing and representing machine
readable meanings. An ontology is the formal specification of the shared conceptual-
ization of a domain of study. In our work surrounding the DCO web portal, an initial
part was the development of domain specific ontologies, and the integration of al-
ready existing ontologies. Our portal adapted the VIVO system as a platform for
metadata management. The VIVO system itself already uses a list of ontologies to
support academic information management. In our work we further extended the
VIVO system by developing a DCO ontology and importing a few other ontologies
such as the PROV Ontology [3] for provenance documentation and DCAT [4] to rep-
resent datasets and data catalogs. Table 1 lists the key ontologies and schemas used in
the web portal.
Table 1 Ontologies and schemas used in the DCO web portal
Name Namespace URL Prefix
Dublin Core Metadata Element Set http://purl.org/dc/elements/1.1/ dc
DCMI Metadata Terms http://purl.org/dc/terms/ dct
VIVO Core http://vivoweb.org/ontology/core# vivo
VIVO Scientific Research Ontology http://vivoweb.org/ontology/scientific- scires
research#
Data Catalog Vocabulary http://www.w3.org/ns/dcat# dcat
Bibliographic Ontology http://purl.org/ontology/bibo/ bibo
Citation Counting and Context Character- http://purl.org/spar/c4o/ c4o
ization Ontology
Citation Typing Ontology http://purl.org/spar/cito/ cito
FRBR-Aligned Bibliographic Ontology http://purl.org/spar/fabio/ fabio
Event Ontology http://purl.org/NET/c4dm/event.owl# event
Friend of a Friend http://xmlns.com/foaf/0.1/ foaf
vCard Ontology http://www.w3.org/2006/vcard/ns# vcard
Geopolitical Ontology http://aims.fao.org/aos/geopolitical.owl# geo
Simple Knowledge Organization System http://www.w3.org/2004/02/skos/core# skos
DCO Ontology http://info.deepcarbon.net/schema# dco
PROV Ontology http://www.w3.org/ns/prov# prov
It should be noted that those ontologies are not separated from each other. Instead,
they are integrated as a whole knowledge graph for representing the various agents,
entities and activities in the DCO scientific community. Ontology reuse and inter-
mapping built the relationships among the components in this knowledge graph. For
example, bibo, c4o, cito and fabio ontologies represent the network of biblio-
graphic and citation information among various types of publications, foaf repre-
sents the network of researchers and organizations, vivo and dco further extend the
inter-connections among those components and other objects such as research topics,
grants, projects, awards, and more. Provenance documentation leverages the W3C
recommendation prov, which represents a high level framework. Classes and proper-
ties in other ontologies, such as dco, vivo and foaf, can be mapped as subclasses
and subproperties of corresponding classes and properties in prov. Moreover, the
knowledge graph can be extended according to real-world needs, especially dco,
which is an ontology created and curated by ourselves for the DCO community.
Fig. 1. A schematic view of the DCO web portal.
2 Annotation and linking to create semantics
With a knowledge graph as the core, the developed DCO web portal consists of four
major parts: Drupal, a content management system used as the main front-end web
portal where users can register, discover and retrieve various types of objects; Handle
System, which is used to assign a persistent and unique identifier to the objects,
known as a DCO-ID; VIVO, the main knowledge store; and CKAN, used for the
storage and archiving of datasets and other media. (Figure 1). Object registration,
discovery and retrieval can all be facilitated through the use of the DCO-ID, similar to
what the Digital Object Identifier (DOI) does for publications.
The functionalities of the web portal enable an individual researcher to record almost
all the components in the life cycle of their research, from funding application, in-
strument deployment, field work planning, data collection, data analysis, meeting
records, publication archival, to project reporting, and more. The portal also allows
researchers who share a common research interest to find, communicate with and
collaborate on research through virtual groups. All instance objects can be annotated
with a list of properties from the corresponding ontologies and can be linked directly
or indirectly to other objects. For example, a journal paper can be tagged with a few
keywords as its topics. One or more of those keywords may also be used to represent
the research interests of a researcher, who might find that paper of interest by search-
ing the keywords, and from the keywords the researcher may in turn find other publi-
cations or researchers within the same domain. Such annotation and interconnections
among instance objects provide a more detailed network about the real world situation
and can expand our understanding of the research to an extent that cannot be reached
by only reading the conventional publications.
3 Identification and persistence of DCO resources
The DCO-ID provides a persistent and unique identifier to all resources in the DCO
web portal. The DCO-ID is similar to the DOI for publications, but it extends the
scope to many more types of objects, including publications, people, organizations,
instruments, datasets, sample collections, keywords, conferences, etc. The environ-
ment of the Web may evolve in the future and the web addresses of the portal and the
various objects registered in it may change. With the DCO-ID, even after 10 or 100
years, one can still find the associated web address of that object and retrieve the in-
formation needed. In this way we can keep a persistent and stable legacy for the activ-
ities and outputs of the DCO community. Fig. 2 shows a DCO publication records,
which has both a DOI and a DCO-ID (shown as a code in the „metadata‟ bar). The
DCO-ID allows users to retrieve more domain specific annotations from the metadata
of the publication in the portal. The records of community, authors, subject areas and
journal shown in Fig. 2 are all hyperlinks and they all have their own DCO-IDs.
Fig. 2 A record in the DCO publication browser.
4 State-of-the-art data stewardship
Data stewardship has a two-fold meaning: data management and data service. Seman-
tic technologies can leverage both parts. The above sections focus more on the data
management side. In our work we also made innovative progress on the data service
side. We are working together with other organizations to advance discovery and
usability of science data as well as other resources. One recent collaboration is with
the output of the Data Type Registry (DTR) working group [5] of the Research Data
Alliance. Each DTR is a self-contained portal for data type registration and curation.
There are some common basic types, which are called „primitives‟ and will be regis-
tered and managed by a central data type registry. This shows a two level hierarchy of
a DTR, one is a list of primitives and the other is the specific data types defined with-
in a DTR. This two-level hierarchy initiated our extension to the DCO ontology. In
our work, the basic data types are classes in the DCO ontology, and the specific data
types are at the instance level, i.e., they are all instances of a newly created class
“dco:DataType” and are part of our knowledge graph but accessible outside of DCO.
Fig. 3 shows the DCO dataset browser, in which the data type is used a facet that can
help users to find dataset of interest.
Fig. 3 Using data type as a facet to retrieve datasets of interest.
Besides dataset curation, we also utilize the latest progress on data citation and sam-
ple collection curation. For example, we created a class „dco:GeoSample‟ in the DCO
ontology which refers to the global initiative International GeoSample Number [6] for
metadata used to annotate geological samples. We also use the metadata schema
DataCite [7] for citation properties of registered datasets in the DCO web portal.
5 Concluding remarks
Our aim for the DCO web portal is to create more than just a data portal, but a
knowledge portal. By using semantic technologies and leveraging state-of-the-art
methods in data stewardship we built a web portal for the DCO community to support
various aspects of their research. The information collected in the portal, from both
the DCO community and extramural data resources, is stored in ways that both hu-
mans and computers can read and understand. A key feature of our portal, as enabled
by the Semantic Web, is the linkage among various registered objects and the flexible
ways to present them. With linked data we are able to create more and better collabo-
rations, find like-minded individuals working on common projects, add data that can
be useful to others, discover tools that can be used to visualize data in new ways, and
make it easier to discover, access, understand and use the data.
References
1. https://deepcarbon.net/
2. Berners-Lee, T., Hendler, J., Lassila, O., 2001. The Semantic Web. Scientific American
284 (5), 34-43.
3. Lebo, T., Sahoo, S., McGuinness, D., 2013. PROV-O: The PROV Ontology. Accessible
at: http://www.w3.org/TR/prov-o/
4. Maali, F., Erickson, J., 2014. Data Catalog Vocabulary (DCAT). Accessible at:
http://www.w3.org/TR/vocab-dcat/
5. https://rd-alliance.org/groups/data-type-registries-wg.html
6. http://www.igsn.org/
7. https://www.datacite.org/node