Upload
consortium-of-universities-for-the-advancement-of-hydrologic-science-inc
View
8
Download
0
Embed Size (px)
DESCRIPTION
2015 CUAHSI Conference on Hydroinformatics
Citation preview
: Sharing Diverse Hydrologic Data Types and
Models as Social Objects within a Hydrologic Information System
Jeffery S. HorsburghMohamed M. Morsy, Anthony M. Castronova, Jonathan L.
Goodall, Tian Gan, Hong Yi, Michael J. Stealey, David G. Tarboton, and the rest of the HydroShare Team
CUAHSI Hydrologic Information SystemEnabling Water Science Data Discovery
But, data and models used by hydrologists are diverse
Time series Geographic rasters Geographic features Multidimensional space/time Model programs Model instances We needed to move beyond time series to a more
general Hydrologic Information System that better supports the data/models we use and the way we work
To the Cloud! Convenient sharing Accessibility anywhere Cross platform Low cost
But Storage, but not much else File formats, content, and
semantics still matter
New Opportunities for Data Sharing and Preservation
CUAHSI
HISSharing hydrologic data
Emerging data repositories Functionality archival/preservation Still very much discipline specific Impact is higher if you choose carefully!
Data repositories do data but not models Model repositories dont support data and
most dont support model instances Most rely on curation of static products
with no real collaborative capabilities
Social Objects
Objects around which social networks form
Jyri Engestrm
What do we want to do?
Easily create a digital instance of a dataset or model (a Resource)
Quickly share it with colleagues (perhaps privately)
Add value through annotation and iteration Describe with metadata Eventuallyshare publicly or formally Publish
Data and models are social objects shared among scientists
Web-based system for advancing data and model sharing
Building on what we learned in developing the CUAHSI HIS to support more diverse data types and models
Our goal: Allowing scientists to create social objects that add value
Why is it hard to enable sharing of hydrologic data and models (Resources)?
Among a host of other technical challenges: Resources may be made up of a single file or
multiple files There may be a hierarchical structure Resources of different types may have different
content data models File formats/hierarchies Syntax
First we needed to define our social objects Resources consisting of hydrologic datasets and models
Then, HydroShare needed a generalized structure within which those objects could be created, stored, described, annotated, and packaged for transmitting over the Internet
HydroShare Resources
Resource = primary unit of digital content Create Share Own Access Filter Discover
We needed to be able to manage all of this functionality consistently across all resource types.
HydroShare Resource Data ModelAn profile of the Open Archives Initiatives Object Reuse and Exchange (OAI-ORE) standard
An XML document
that encodes the
description of a Resource
and the Aggregation
A list of all of the objects/files
aggregated within the resource
A file that is part of a Resource
OAI-ORE = A general standard for description and exchange of aggregations of web resources
Simple Example: Hydrologic Time Series
Formal semantic terms are used to express relationships among objects:o The Resource Map document describes the Aggregationo The Aggregation aggregates the content fileo Expressed as RDF triples
A computer can learn the structure of a Resource by reading its Resource Map document
Resource Metadata: Dublin CoreCommon to Every Resource
Resource Content Data Models
Resource Metadata: ExtendedSpecific Elements for Each Resource Type
Packaging Resources
How to store resources on disk?
What do you get when you download a resource?
Packaging Resources for Storage and Transfer
Bag-It! A hierarchical file packaging format for storage and
transfer of arbitrary digital content Storage on disk and serialization for download
Model and Model Instance Resources
Public and Private Sharing
Set as Public or Private
Choose a license
Decide who has access and what permissions they
have
Ratings and Comments+1 a
Resource
Start a Conversation
+1 a Comment
Receive notifications
What if?
Dataset deposited in HydroShare
Paper using theDataset is published
Dataset annotated by HydroShare users
Dataset synthesized and leads to another publication
TimeInf
orm
atio
n Co
nten
t of D
ata
and
Met
adat
a
Summary Hydrologic datasets and models are social objects HydroShares Resource Data Model enables us to
consistently handle diverse Resource types Machine and human interpretable Resource content data models add structure to
known Resource types Resource Data Model = Container Resource Content Data Model = Whats in the container
Storage on disk, access control, transport over the Internet, and cataloging are consistent for all Resource types
Web Resources
HydroShare system: http://www.hydroshare.org
HydroShare project website: http://hydroshare.cuahsi.org
HydroShare GitHub repositories: https://github.com/hydroshare/
Questions?Support:ACI 1148453ACI 1148090
: Sharing Diverse Hydrologic Data Types and Models as Social Objects within a Hydrologic Information SystemCUAHSI Hydrologic Information SystemEnabling Water Science Data DiscoveryBut, data and models used by hydrologists are diverseTo the Cloud!New Opportunities for Data Sharing and PreservationSocial ObjectsWhat do we want to do?Slide Number 8Why is it hard to enable sharing of hydrologic data and models (Resources)?Slide Number 10HydroShare ResourcesHydroShare Resource Data ModelSimple Example: Hydrologic Time SeriesResource Metadata: Dublin CoreCommon to Every ResourceResource Content Data ModelsResource Metadata: ExtendedSpecific Elements for Each Resource TypePackaging ResourcesPackaging Resources for Storage and TransferModel and Model Instance ResourcesPublic and Private SharingRatings and CommentsWhat if?Summary Web Resources