The session detailed the state of both and more importantly, their relationship.
Karen Siderlis on the Geospatial Platform
Geospatial Platform Definition: “The Geospatial Platform will be a managed portfolio of common geospatial data, services, and applications contributed and administered by authoritative sources and hosted on a shared infrastructure, for use by government agencies and partners to meet their mission needs and the broader needs of the Nation.”
Target customers: “users,” geo professionals, managers, lay public. Right now there’s a greater focus on geospatial professionals.
Providers: not only federal gov data, ngo, academia, state and local, commercial data, VGI
One more version of the roadmap coming, then will focus on implementation strategies. (V3 is on website.)
The map on the site (Deepwater Horizon) is “what it might look like,” this will be replaced by the “real one.” (CNN coverage and GCN award. NOAA took the lead on this map. This model might work in the future, a shared leadership model.)
Jerry Johnston, GIO EPA on Relationship Between Platforms
The relationship between the Geospatial Platform and data.gov is evolving. Data.gov is about getting data to people. Geoplatform’s focus is about agencies carrying out their mission (which may have a public face). It’s not necessarily about providing data.
Cross agency team lead by GSA started work in March 2009. Data.gov was up and running May of that year. 47 datasets at launch, with 2 million hits that day. Now 300,000+ datasets. How are people using this data? How many apps are available using its data? Many are mapped based. His favorites:
Flight delays (based on FAA data).
Community health (First Lady’s efforts against obesity)
Employment Market Explorer (jobs)
Datamasher.org (make mashups yourself)
Now the team doesn’t measure datasets but how they are used. 90% of the data is geospatial data because of Geospatial One-Stop “input.” But also, people are interested in maps.
About six months ago we added a mapping preview capability - to “explore data” for fitness of use, not for making mashups. It’s not slick, it’s dots on a map. But in the backend, the data are in shapefiles, not Web services. Data.gov actually grabs shapefiles and makes a Web service which is visualized. We hope to get all the data as services, so this is a workaround. It adds a lot of value.
One things that’s missing: spatial search. Right now only keyword based; it’s getting better, too. Now it’s using Microsoft Bing indexer. We are working on bounding box search. That will leverage GOS work. We’ll decommission GOS in favor of Data.gov.
Also coming: search place names, gazetteer search, metadata harvesting. Now we are putting web services into the catalog.
What about other data in GOS, non-federal data? There could be two catalogs - one Fed and one other, but data.gov would have just Fed. At Geosplatform there would be access to both. Or we could put other data into the federal catalog. The reason for the restriction? Data quality. Maybe we need to revisit that decision. “We will solve this problem.”
We think data.gov could be a great resource for the Geospatial Platform. We want to have people do more with data when on data.gov - analysis, sharing, etc. THat is, increase “stickiness.” To get there there’s an RFP for “dataset hosting” which is aimed at getting data and easing the agency challenge of putting out data for the public. (Socrata got a contract). This would be a “place to put data” and some simple data exploration. Maybe there could be a geospatial data hosting capability at data.gov.
Q and A
Data by FIPS code in Data.gov?
(JJ) Yes, lots.
LOC: Easy way to link to controlled [metadata] vocabulary for agencies?
(JJ) Yes, but its cludgey. Can’t do it “meaningfully” but issue is raised. FGDC data works.
Stop gap: just a link?
(JJ) Yes, that’s a great idea.
How about a thesaurus for that vocabulary?
(JJ) EPA is looking at something like that. That’s a science problem, a government-wide program. We need a federal focus on this problem.
No longer separation between raw and geodata coming, but the standards would be the same.
(JJ) User would not see it.
GOS retirement. Will there be a stepwise approach to the move?
(JJ) We are looking at contract for that for July. But we need to coordinate/communicate with each agency on the steps. Expect May/June communication on the matter. Hopefully the transition will be transparent to the user.
Agencies uploading summarized or raw data?
(JJ) Up to the agency. OMB wanted “a lot of data.”
(JJ) We control for “mosaic effect” - by combining datasets could get something you shouldn’t have. Someone reviews that for national security potential.
What analysis is coming?
(JJ) Socrata has sorting, grouping by field, sum/avg. The interesting thing is queries that can be made as a new dataset (KML, text file).
Bill Burgess: Comments were over Dec 31. Why are you asking for more?
(KS) Goal was to close by then, but extended to end of January 2011. This conference is a venue for more comment. Other things going on, review by Gartner.
Bill Burgess: Turnout in IdeaScale has been lame. We want to get more interest. Highest vote is 35! Most folks won’t read the road map.
(JJ) Data.gov had good return (White House push) but Geospatial Platform did not. We need to get something out there and then get a response.
How do we make it easier to identify data to share?
(JJ) Crawler for data.gov to capture every file on every agency website to make a report back to agency was proposed. Lots of resistance.
Where will Marketplace (GOS) go?
(JJ) Somewhere. It’s important.
Data.gov needs a metadata tool to reconstruct it for old data.
(JJ) EPA has such a tool, maybe.