“Leave geospatial data where it lives
and transform it into GeoJSON, CSV, KML, a Shapefile,
or a Feature Service dynamically.”

Earlier this month, we took our cloud analytics show on the road to Esri‘s partner conference and developer summit, which is a week devoted to the business and ecosystem around this market-leading geographic information system software provider.

Esri’s flagship offering is ArcGIS, which comes in the following versions:

  • desktop version targeted at professional cartography and map-centric enterprise apps
  • a server version for storing and indexing spatial data
  • the relatively new ArcGIS Online

IBM and Esri have a strong, venerable partnership going back to the 1990s. At this year’s conference, our IBM team was there to highlight the ways you can integrate dashDB and Cloudant with Esri technology. dashDB, a new offering that’s a close sibling of DB2, takes advantage of the time-tested DB2 spatial integration with Esri. You can use dashDB as an enterprise data warehouse, but still connect to it from ArcGIS (desktop or online) for high-end mapping functionality. From an organizational perspective, this is powerful since it lets mapping and GIS users work seamlessly within the traditional corporate IT world. Companies already use dashDB as an EDW or “data lake”, so the ability to do a File -> Open on dashDB data from within an Esri tool opens up that information to GIS-centric solutions. For more details, see this slide presentation by IBM’s John Park:

and check out these tutorials:

Since I’m particularly interested in open data and open source, I was impressed by the latest advances to ArcGIS Open Data. Esri-hosted ArcGIS Open Data gives you a quick way to set up public-facing websites where people can easily find and download your open data in a variety of open formats. It’s powered by Koop, an open source geospatial ETL engine that aspires to let you, “Leave geospatial data where it lives and transform it into GeoJSON, CSV, KML, a Shapefile, or a Feature Service dynamically.”

Koopred numbers image by Dave Bleasdale

Koop is a product of the Washington, DC Esri R&D team, originally conceived by Chris Helm, when he worked there, and now maintained by Daniel Fenton. Here at IBM, Norman Barker leads our effort to build a “Koop provider” for the Cloudant NoSQL JSON document store, so that any data stored as GeoJSON in Cloudant can be accessed through Koop. While Cloudant already supports advanced spatial queries and indexing for extremely fast access, putting a Koop interface in front of Cloudant lets developers re-use Esri style queries and data format skills on data in Cloudant.


This has powerful implications–not only for online mapping–but also for analysis, as I discussed in a Cloudant-Koop lightning talk at the conference. Join us in developing the Cloudant-Koop provider on GitHub.

dodgeballThe week’s activities culminated with a wildly popular and intense dodgeball tournament, which IBM sponsored. Our team made it out of the opening round, but quickly fell in round 2. I clearly need to work on my fast-twitch muscles to have a better showing next year. But we all had a good time and look forward to using new relationships and technologies to cook up even better online mapping and spatial analysis solutions for you in the future.

Join The Discussion

Your email address will not be published. Required fields are marked *