A presentation of the underlying motivations and institutional context behind GeoNode, some of its major design decisions, and unresolved challenges for its sustainability.
I gave this talk at UC Berkeley School of Information's research seminar on Information and Communication Technology for Development (ICTD).
Much of the material comes from an older presentation I wrote with Rolando Peñate.
1. GeoNode Motivations, Design, and Challenges
Sebastian Benthall
UC Berkeley School of Information
ICTD Seminar
(Based on a presentation written with
Rolando Peñate
OpenGeo)
4. Spatial Data Infrastructure (SDI)
“[Spatial Data Infrastructure] provides a basis for spatial
data discovery, evaluation, and application for users and
providers within all levels of government, the commercial
sector, the non-profit sector, academia and by citizens in
general.”
– SDI Cookbook
5. The theory of SDI
developed before
we learned what was
possible with the Internet
17. What is GeoNode?
GeoNode is a spatial data infrastructure
It focuses on data, then users, then metadata.
Data upload, sharing, cartography, user profiles, dynamic metadata
generation, and more.
18. What is GeoNode?
GeoNode builds on open source geospatial projects like
GeoExt, OpenLayers, GeoWebCache
GeoServer, GeoNetwork, and PostGIS
with application functionality built on Django.
19. GeoNode Vision
⇒
GeoNode Involvement
⇒
GeoNode Community
20. GeoNode Vision
⇒
GeoNode Involvement
⇒
GeoNode Community
How did this happen?
21. Can the lessons learned can
help other ICTD projects?
A case study GeoNode
sheds light on international disaster
reduction efforts.
22. Disaster Risk Modeling 101
Risk
• Used for (busted stuff)
determining =
development
Hazard
investments (boom)
• Once were a mess x
• Now standardizing: Exposure
(stuff)
x
Vulnerability
(bust per boom)
23. GeoNode History
The World Bank had a problem:
Disaster risk modeling requires lots of data
Central American Probabilistic Risk Assessment (CAPRA) initiative
needed participating agencies across various governments to share
data
Top-down approaches didn't work
Needed to work bottom-up
24. GeoNode History
The World Bank had a problem:
Costly proprietary GIS solutions are a burden to developing nations
The Bank wanted to build local capacity around financially
sustainable software
Smart folks within the Bank turned to open source geospatial
software
25. GeoNode Vision
OpenGeo had an idea for a solution:
The Bank provided the perfect use case for OpenGeo's vision for open
source architectures of participation in geospatial
Providing freely available web-based tools could be a great way to
collect and share data.
GeoNode was born.
26. GeoNode Involvement
Traditional SDIs have typically been designed by 'experts' with
abstract needs in mind—hence a focus on metadata.
GeoNode is being designed in response to the needs and
concerns of institutional partners as they implement real-world
projects—hence a focus on data and users.
33. Open Data Optimism
Features like
User reputation
Organizational endorsement
Flexible security
address data quality concerns
34. Open Data Optimism
GeoNode supports
the continuum
of openness with a common platform
for institutional GIS and neogeography
35. GeoNode Involvement
GeoNode seeks to unify data management across
organizations.
Thus many different organizations have reason to get involved.
The opportunity and challenge is effective collaboration.
36. GeoNode Involvement
As more organizations got involved, development had to
decentralize.
Not just a single team within OpenGeo, but a larger community
37. How do we continue growth when
vision and development are decentralized?
38. How do we continue growth when
vision and development are decentralized?
40. But how do we get institutions to get their employees
to participate in the open community?
Need to align broader visions, including...
41. Disaster Reduction
• Australia-Indonesia Facility for Disaster Reduction
• Geoscience Australia
• Global Earthquake Model
• Global Facility for Disaster Risk Reduction
• Secretariat of the Pacific
are mapping infrastructure in developing nations, performing
disaster modelling, etc. using GeoNode.
42.
43.
44. Academic
MapStor Foundation and Harvard's WorldMap seek to collect and
share data across disciplines and institutions using GeoNode.
45.
46. Spatial Marketplaces
The Australia–New Zealand Spatial Marketplace seeks to increase
data availability in the South Pacific by creating an online
marketplace built on GeoNode and open to all.
47. Community
The World Bank's vision was the
collaboration of many institutions and governments
around common goals of data management
48. Community
As a result,
many organizations are involved
in building and extending GeoNode
49. Community
How can we keep these efforts coherent, not divergent?
Efficient, not redundant?
51. OpenGeo
• Benefits from contributions back to core software
• Has led effort to coordinate between institutions
o easier management and development
o stronger open source communities
52. Our task has been to
scale up open source development
practices to large institutions
53. Roadmapping Summit May 2011
• Explicit transition to open source community model
o Established a proper Project Steering Committee
o Passed policies for contributions and code review
• Official decentralization from OpenGeo's core team
• Identified common development goals
64. Outcomes
• "Rock Solid" 1.1
• People entered the summit to big ideas to impress their
bosses
• People left having committed resources to docs, bug
fixes, and other work necessary to keep the project
running.
65. Outcomes
• Framework for future improvements
• We have principled roadmap for the software with real
institutional backing
• We know who to call when we have the resources
71. Remaining (technical) challenges
Is the dream of a
secure federated data network
(both spatial and social)
realistic?
This ties into questions of federated social networking.
72. Remaining (research) challenges
This perspective on GeoNode is from
offices in New York City and Washington, DC
What does it look like in the countries
where it is being deployed
I’d like to talk a bit about SDI's. Specifically what I think is the last step in building the dream of INSPIRE. If we're to look at this one definition of SDI INSPIRE has already done most every thing needed - it's written in to law, there are standards, everyone's agreed and it's moving forward. But I have a worry:
Traditional SDI : Metadata catalog, sometimes provides access to data, rarely incorporates users. GeoNode : Data first, then users, then metadata is derived where possible. The focus on data means we strive to make it easy to upload data and have it all served up automatically in all available formats. It's easy to access with a URL to a page with the data and the metadata. It's the actual data so you can see it's rough distribution, and you can easily make a map of it, combining with other data. It has links to the services, but also direct links to formats people may want from the services. I always wonder how I am supposed to give someone a link to a layer in a WMS. Is it the capabilities document? A top level tile? Or a catalog response in ISO19115 that gives me the service links? Then we bring in users. Every data set is associated with a user, and that user's profile information populates the ISO metadata fields, so you don't have to fill it out every time. From users and data we get to metadata. We use GeoNetwork as the CS-W engine but we try to populate as much as possible from the user who uploaded the data (via the user profile) and the data itself (by extracting bounding box, etc). Finally we make 'maps' a top level concept that people can share and use, so you can more easily explore data, and from there also derive more metadata (eg, if 5 maps that all use the same layer have a tag 'fire' then perhaps that layer is about fire). Another thing GeoNode does is always coordinate the right links between the capabilities documents of the CSW and WMS/WFS/WCS, so each refers to the other, and also contains the same abstract, keywords, etc.
In many ways, GeoNode is the product of several very abstract ideas voiced by Stuart Gill of the World Bank and Chris Holmes of OpenGeo. Top-down == mandating policy.
The Bank has an interest in promoting IT infrastructure in the regions where it works to promote development Can pay for proprietary software in an initiative, but that sticks countries with costs after the Bank leaves
was started to encourage bottom-up data sharing while limiting costly proprietary GIS installations to only those agencies that really needed them.
Global Earthquake Model
SOPAC
Harvard WorldMap
Emphasis on public participation—architectures of participation, yay!
Whereas we had approximated open source community-based developed as a team internal to opengeo, we used the opportunity to transition to an official, public community model. -- though one that is still over-represented by OpenGeo and World Bank developers thus far.
RP: Notes for this?
RP: Notes for this?
Trevor as an example of turnover at AIFDR. He has largely spearheaded the idea of propagating GeoNodes across government agencies in Indonesia but his term is almost up.