OpenEarth: a 4th paradigm workflow for marine and coastal science

Post on 18-Jul-2015

296 views 0 download

transcript

`

Dutch Data Award for beta sciences 2012

Feb 28th 2013, SRIE 2013

Need for marine and coastal science/engineering http://nupedailynews.com/wp-content/uploads/2013/01/telescope..jpg

http://chromblog.thermoscientific.com/Portals/49739/images/ lims for biobanking1.png

http://www.eiroforum.org/media/photo_galleries/cern/cern-02l.jpg

http://oeatech.net/wp-content/uploads/2011/03/RADARSAT2-satellite.jpg

suspended mud particle

64 μm

6400 km: North Atlantic Oscillation (NAO)

gap 10-6 > 10+6

Water scarcity Flood protection Water pollution

http://upload.wikimedia.org/wikipedia/commons/2/22/ Da_Vinci_Vitruve_Luc_Viatour.jpg

Marine/coastal scales

OpenEarth is a workflow

philosophy technology

community

OpenEarth philosophy

Open up & collaborate

philosophy Non-zero sum game: 1 + 1 = 3 Tragedy of the Commons OpenEarth wants to keep sharing going

OpenEarth philosophy

Open up & collaborate

community

OpenEarth philosophy

Open up & collaborate

• all collaborating is not enough • chaos • coordination needed • but no overall boss • we need smooth workflow • like wikipedia

OpenEarth philosophy

technology

Open up & collaborate

check this out

OpenEarth community: social media & hands-on

philosophy technology

community

LinkedIn group 269 members

Sprint sessions: hands-on outreach & feed-back every year prior to Netherlands Centrum voor Kustonderzoek (NCK) days

OpenEarth philosophy: universal principles

philosophy Technology principles

community

1) data & tools are one: database alone is useless

raw data: volts counts

+ software

data product =

https://earth.esa.int/handbooks/meris/aux-files/image047.jpg http://logfurniturehowto.com/wp-content/uploads/2012/07/ Types-Of-Screwdrivers-And-Their-Uses.jpg

ESA/NASA raw satellite volts

ESA/NASA useful product physical units

2) good enough to share, if good enough to publish transparency to combat sloppiness, mistakes & fraud. No standardization yet.

http://adaywithsfe.files.wordpress.com/2012/09/eazie5.jpg

http://static.nl.groupon-content.net/25/05/1333050000525.jpg

http://cdn.chud.com/a/af/afdcafa0_2601_vegetable_lo_mein.jpeg

http://farm4.static.flickr.com/3330/3599399622_839ec0cc66.jpg http://blog.gjvanbussel.nl/wp-content/uploads/2011/09/Stapel.jpg

3) Quality is a version control process

• Sharing will generate feed-back: version control becomes inevitable • update your own work • allow altered copies to proliferate

• Version control is common practice for • ISO 9001: doc, xls, pdf • Toyota Kaizen: keep on improving • Software engineering: windows update • … science is lacking behind anyway

http://www.knups.nl/leuk_plaatje/56/Helaas_pindakaas.html http://www.food-info.net/nl/national/ww-pindakaas_clip_image001.jpg

OpenEarth technology: a pragmatic choice

philosophy Technology: an open software stack

community

Off-The-Shelf (OTS) vs off-the-drawing board

http://www.bnr.nl/incoming/621827-1301/fyra-578.jpg/ALTERNATES/i/Fyra-578.jpg

http://1.bp.blogspot.com/-Z8vfA9SmoOg/T84rGCE_XaI/AAAAAAAABLM/Wp2nXaTvrEc/s1600/Amsterdam+086.jpg

We seek an open source software stack that can serve our needs today. We do embrace new technology, but will not allow it to delay us from proper data management now.

Cloud technology: operational de facto web-standards

tailored data > OGC WxS …

> ISO SQL- PostGIS > OGC netCDF- CF OPeNDAP

SubVersion …

> OGC KML …

… graphics of data

standard data

raw data

catalogue of data

work done on server

work done on client

what where when who why how … data URLs

smart phone & tablet users scientists professionals

Subversion: version control on data and tools

Subversion: version control on data and tools

# commits

Time in quarters

Web services to data subsets: linking with DOI

All data in Google Earth: DataTube weather models

10 km resolution

satellite data

1km resolution

Dutch digital elevation maps

100 m resolution

scale of interest: km/yr

coastal bathymetry

10 m resolution

dune profiles

1m resolution

Pro

ces

scal

e: k

m/y

r

scale of interest: km/yr

Data scarcity > data abundance: data scientist most sexy job 21st century 1. 1000’s yrs: Empirical Archimedes ea. 2. 100’s yrs: Theoretical Newton ea. 3. 10’s yrs: Computational Neumann ea. 4. Now 4th paradigm by Microsoft Research

Portal to filter: deal with complaints overwhelmed users

Questions