owned this note
owned this note
Published
Linked with GitHub
# 2021-02-04 DataONE Community Call
[![hackmd-github-sync-badge](https://hackmd.io/npMXMuvsTAuXHC8-mjomkQ/badge)](https://hackmd.io/npMXMuvsTAuXHC8-mjomkQ)
**Topic**: How can DataONE interact with emerging networks
**Time**
* Pacific Time (US): 16:00 (02/04/2021)
* Mountain (US): 17:00 (02/04)
* Central (US): 18:00 (02/04)
* Eastern (US): 19:00 (02/04)
* UTC: 00:00 (02/05)
* Sydney (Australia): 11:00 (02/05)
**Description**: Many of us participate in multiple cyberinfrastructure, disciplinary, and
professional networks as part of our daily work. While these networks frequently provide
complementary resources and capabilities there are cases where there might be duplicative
efforts or capabilities that coordination between networks might help reduce (if
appropriate). This community call will bring together panelists from multiple
networks that have some degree of existing connectivity with the DataONE network and
community and start a discussion focused on opportunities for collaborative and coordinated
activities that can efficiently meet and provide cost-effective solutions for shared
community needs while contributing to increased sustainability of network capabilities.
**Invited Panelists**:
* Kerstin Lehnert
* Margaret O'Brien
* Mike Daniels
* Pip Bricher
* Matt Jones
**Attendees:** (name, affiliation, email)
* Amber Budden, DataONE/UCSB (aebudden@nceas.ucsb.edu)
* Matt Jones, DataONE/UCSB (jones@nceas.ucsb.edu)
* Kerstin Lehnert, IEDA / Lamont
* Pip Bricher
* Margaret O'Brien EDI, LTER, MBON (margaret.obrien@ucsb.edu)
* Karl Benedict
* Mike Daniels NCAR/Ronin/EarthCube
* Susan Shingledecker - ESIP - susanshingledecker@esipfed.org
* Denise Hills, Geo. Survey of Alabama/Ronin (denise.j.hills@gmail.com)
* Joan Damerow, ESS-DIVE LBNL, JoanDamerow@lbl.gov
* Brian Westra
* Corinna Gries, Environmental Data Initiative
* Dave Vieglais
* Erin McLean, Arctic Data Center, mclean@nceas.ucsb.edu
* Mark Servilla
* Stefanie Butland, rOpenSci, stefanie@ropensci.org
* Stevan Earl
* Steve Aulenbach, USGS, saulenbach@usgs.gov
* Thomas Thelen
* William Michener
**High-level Agenda**
1. Introduction (5 min - Karl)
1. Introductions from the panelists (10-20 min)
2. Discussion of next steps to move forward on near-term opportunities for network collaboration (Open discussion about connectivity and collaboration opportunities (10 min - Karl)
**Potential Framing questions for discussion - add your own**
* Where are the productive boundaries and connections between the infrastructure developed and maintained by different Earth science networks?
* What are the benefits and costs of shared infrastructure? Is there more than infrastructure to share?
* What interoperability models can be used to build connections between networks?
* How can we leverage existing network connections - shared individual participation, overlapping organizational memberships, common sponsors, common use of interoperabilty standards?
* What are the challenges and opportunities for sustaining our networks through shared/collaborative work?
* Crawling vs. harvesting as models for federation or connectivity
**Shared Notes**
* Overview / Intros by Karl
* Kerstin Lehnert - CZ Network is a new effort within a new program within NSF. Network of 9 clusters that have different scientific foci. Each cluster is a mini network. KL involved in building data infrastructure for CZ with Hydroshare - submission and access for PIs. has already approach DataONE due to DataONE hosting a lot of data the CZ would like to integrate with. No need to rebuild somehting that exists. The other network is samples - physical onjects that many studies focus on that generate data. Internet of samples and RCN 'Sampling Nature' just funded (but not started). Connection btn DataONE and sample networks provides new opportunities. Shared infrastructure - Council of Data Facilities. CDF, formed through Earth Cube, forum for discussion. has had a 2 yr WG exploring shared infrastructure. Also, OneGeochemisty: International group trying to build data standards for lab analytical data. dataONE network would bring a lot of structure to a new effort trying to build data standards, best practices etc.
* Margaret O'Brien - Information manager with EDI. Domain repository but do accept any data witha focus on environmental data, and long term. Focussed on infrastructure for the repo itself and also metadata standards. manage 2 DataONE members. EDI approach leverages decades of experience and collaboraiton. Came out of LTER network. Repo software originated there and has continued to develop. History means they have an understanding fo the needs of researcher. In 2016 EDI added data repository support for all environmental data, so although EDi seems young, it has a long history. Also affiliated with MBON. 2 ESIP activities - Sustainable Data Management cluster (https://wiki.esipfed.org/Sustainable_Data_Management, see papers - Gries et al, DSJ) is focused on repositories working in similar ways. current project is to align FAIR/TRUST/CARE principles and create general repo guidelines. A new ESIP cluster for biological data (organismal) arising from interest of NOAA/NASA as related to the BON.
* Mike Daniels - NCAR, 30+ years :) federally funded research dev center, operating obs equipment. Been long term participant in ESIP and Earth Cube. Notcied a shift in early 2010s when OSTP memo came out around open data. Been involved with Earth cube on linked data. Wants to highlight: schema.org is a good model. Got ead funding through earth Cube, community building through ESIP. Very light way to provide some great functionality adopted by goggle and yahoo (?)
* Mike: Center for Atmospheric Research – federally funded research center that operates major equipment and faciliaities. NCAR – aircraft, radar, etc. Observational data. Involved with ESIP as part of his work, and also EarthCube. Shift in early 2010s in open access and open software. Much needed emphasis on CI and data. Linked data and real time data, what do FAIR prinicples mean for real time data, kind of like the samples Kersten works on. Good model for building interop between netowrks - schema,org work has been getting seed funding from Earth Cube, community building with ESIP, low overhead way to provide search capabities that are also supported by a private company (Google).
* Pip Bricher - data officer for southern ocean observing system. work sepcifically in science space (people that collection observations and scientists that will use those obs). Multidisciplinary, multinational. Most oceans have a small number of countries that are heavily engaged. SOOS have 36 nations.
* Pip - Federated data discovery in the polar community (POLDER). Data Officer for the SOOS - multidisciplinary work between people who collect those data nd sthe scintests who use it. Multinational as well - 36 nations - unusal since most oceans are more bounded. Anything connected to the southern ocean. Cleaned up, curated data gets serve through soosmap.aq - built by the physics group, then adding in other datasets like marine mammals with CTD loggers, microplastics data, krill data, etc. EMODnet have also built an equivalent portal for the Arctic. Always going to be a long tail of data that is discoverable via metadata. GCMD provided a metadata search tool for Antarctic data. Oceanographic programs haven't put their data in GCMD - offers a partial view of what they're interested in. Got together with Arctic Data Committee and SCADM to facilitate better metadata search. People sharing their metadata in a way to make it more accessible via a central portal. Schema.org is the way they're making that happen. Filled in a matrix of who is harbvesting what from whom, and what tech they're using from that - created a nightmare of a web. Lots of players working in this space. This is the current state, needed to define that for future movement. github.com/POLDER-Crew & other efforts with oceanographic data are happening as well. And you can find out more about POLDER here: http://www.soos.aq/activities/task-teams/polder And you can find out more about POLDER here: http://www.soos.aq/activities/task-teams/polder
* Matt Jones - DataONE is fundamentally about repositories and connecting networks. Original idea: a shared programming interface - both reading and writing - better that haveing everyone need to build their own. As a community, build interoperable tools. Number of repositories that share infrastructure but with heterogeneity in the network too. Been working on metadata quality service which can support curaiton team in data processing. Working on harvesting citations (scythe package). DataONE networks is dependent on being able to harvest a catalog across the repositories, which is dependent on metadata vocabulairies. Some groups find a challenge to provide comprehensive metadata cards to integration. Schema.org really useful in this regard. DataONE can harvest any group and build a network based on schema.org. Currently only harvesting repos that request to be a member of the network. Could harvest without permissions? AKin to google. Also have portals that support communities in making their data are discoverable, pulling together data from across the network. All data will show in a single simple search via the portal. This portal functionality enables us to build arbitary portals around groups or collections. That comes with metadata assessment (e.g. FAIR profile), citation lists, aggregared metrics. Demoing this with the POLDER group, EarthCube. Provides collective view of data holdings.really focussed on tryng to provide services to he network. What would be helpful for repos within their community networks? Impetus for this call.
* Matt - DataONE is fundamentally about connecting netwowrks of repositories and trying to build a federated network. Having a shared network would save a lot of us time and effort. Reading and writing data with the same tools that would be able to have that interoperability built in. Shared API has been very successful, a lot of repos share the same underlying structure, while acknowleidgng that there's massivve heterogenity in data and data management practices. Here to try to help community get a handle on that hetereogeneity. Prodced shared services that are valuable to the repos in the netowrk, eg FAIR evaluation for all the datasets in the network, useful to curatorial teams to figure out where improvements can be made. Also work on data citation services and tools like scythe to harvest citations. Netowrk services are critical to what DataONE does, depnds on being able to work axcross the network. Done a lot of work on crosswalking btwn various metadata standards. Barrier to entry for a lot of new repos. Movement towards schema is a good thing - getting a core set of parameters is a valuable services. We work with them to find the set of guidleines to apply across the network. Can harvest any group that's interested in being harvested. Right now, only harvesting repos that request to be members of the network, but we could look for more sites that are using the schema.org. We can also build portals that allow people to curate collections of data relevant to a particular community. EX: data portal from Toolik Field Station. Have many other datasets across repositories, and they wanted all their data to show up in one search. This functionality allows us to also build collections for different groups - for example, we could build an Arctic Data Federation portal prototype. FAIR profiles would be available for the whole collection, as well as the citations available. Done something similar with the EarthCube data. Straightforward for us to build these portals with custom serach for groups that want that collective view of their data holdings. What is the set of services that would be helpful for repos both for their community and within the larger community of repos?
**Discussion**
* KL - important to find out what others are doing. networking is critical so that we're able to share best practices and other ideas. Big benefit of establishing a network of networks.
* PB - POLDER demo is largely NA / Arctic, due to nature of DataONE. Hard to get funding with 36 nations given that folks need to fit in with what their local priorities are. What can we do? Current thought on their side is schema.org, but is there another way to do it?
* MJ: Goal at DataONE is to not make them onerous to maintain. Would be nice if any group could create a portal for small $$ - but folks can get together and hopefully join forces. DataONE could help build a sustainability model around individual groups each contributing a small amount. Zero isn't sustainable. NA emphasis exists. Have a few international partners - thinking about turning on the spigot and harvesting data from groups around the world. Reflects funding from NSF but infrastructure doesn't limit ability to work internationally. DataONE can handle around a dozen or more common metadata vocaulaties.
* PB - if 20+ groups were pushing schema.org would that help DataONE with scaling up?
* MJ: still pushing for folks to be able to join for free but have others join as plus members for added services for sustainability.
* KL: See a opportunity for DataONE to work across agencies and do integration - US siloed when it comes to data management. Thinking about CZ - one of the efforts close to CZ activity is ESS-DIVE (DoE). Attractive to integrate CZ data with a group from a different agency. USGS another large holder of data, NASA. Can bring new fuding into the organiziton. USGS, NASA
* MJ: Some of those groups are already part of the network - NCEI, USGS, ESS-DIVE, ARM. However agencies are huge and only a small amount is reflected in the network. How to get entree into the agencies.
* KL: Would be good maybe to start with the data closest to DataONE - the environmental data.
* MD: Bringing in smaller groups, that don't have resources to even implement schema.org, this is a big plus for DataONE. Controlled vocaublaries - trying to push google to develop the VariablesMeasured part of schema.org. These are focussed on dicsovery but interoperability is also critical.
* KB - is there more than infrastructure to share? Metadata is the fuel that keeps us moving forward. Increased focus on schema as a more lightweight/accessible standard. Emergence of more speicifc profiles as a way to provide enhancements to it. Will it lead to fragmentation or lead to multilayered information that fits in the schema envelope?
* PB: Within POLDER group been working on matrix of metadata for discovery depending on what groups want in terms of search capability. Trying to stay aligned with community norms / best practices to cross fertilize and avod siloing. Metadata informed by how comprehensive you want to go. Some data centers have large big internal processors where they might be able to layer schema.org on top. For others, schema.org will be the only implementation
* MJ: One fo the major costs for onoarding is trying to resolve different dialects. About half a dozen different parsing for (XXX metadata field - didn't catch it). Also simple right now, schema.org has capability of being expanded for different needs. Qualitatively different from yet another XML based monolithic metadata standard.
* KL: DataONE has a system, and implementation can text these things. DataONE can use what it has as a test bed.
* KB: This is an area where it might be good to think about how to leverage capabilities developed in our networks, leverage connections (human and infrastructure).
* MD: Another emerging thing addressing the IAR (of FAIR) are notebooks. Makes codes and methods more transparent and sharable, translate of methods in research.
* MJ: DataONE has been participating in WholeTale. Projects like that help in the publication of complex datasets, information, workflows. Lot of potential for growth. Repos need to be engaged in that conversation, help build formal linkages to data repositories.
* KB: Steamline production of that basic documentation to facilitate research
* SS: Quick intro - new ED of ESIP. Connecting people and communities - best ways to contribute to the conversation and bring people together. Reach out!
* MO: Sustainable DM group at ESIP, focus was on repos working together. group still exist.
* KL: Think should follow up on CDF link.
* MJ: Thinking a lot about relationship between ESIP, DataONE, RDA etc. Participate in many ESIP cluster. ESIP largely US focus, RDA includes more international. Have talked in the past with Erin about how ESIP provides the community fabric around which this is based. DataONE focusses more on ground up implementation. Need synergy -take ideas from ESIP and move into practice. How to better build relationship.
* SS: On NA focus, some constraints around funding sources. Increased reach in Australia / NZ.
* KB: Convene specific folks - DataONE, RDA, esip clusters, CDF, bring folks together to look how they can more efficiently interact with each other or partion where efforts are being focused. + others for targeted discussion & continue what we've started here.
* MJ: Also consider European networks that have federated around other networks - embraced RDA but not ESIP.
* KL: Was a network at AGU 19 - AUS, EU, Earthcube, IRIS, etc. Tried to get international networks together. Earth science focused but could be extended.
* MJ: NEON Partnering with a bunch of international networks doing environmental data monitoring: GERI: https://www.neonscience.org/impact/observatory-blog/global-ecosystem-research-infrastructure-geri-agreement-signed
* KB: Next steps - Karl to reach out and perhaps form a small group to organize another meeting around these topics. Bring in other networks to the call.
**Links/notes from chat**
That polar harvesting graph is here: http://staging.arctic-data-ecosystem.apps.nsidc.org/harvest
And you can find out more about POLDER here: http://www.soos.aq/activities/task-teams/polder
And the POLDER GitHub repo is just firing up now but is here: https://github.com/POLDER-Crew