HHypermap: A Platform to Enable Geospatial Search

Status:
Accepted

There is currently no comprehensive search capability for online spatial information, a critical body of data which now represents an uncurated, “dark web” of data services and data files. One could argue this domain is not even a web since most of it does not link out and most is not linked-to from elsewhere.

Though there are millions of geo-services and datasets just a click away, there is no way to discover most of them. It is difficult to estimate the opportunity cost of having no general purpose geospatial search capability. Imagine not being able to find a web page. That was the way it was for HTML in 1995, and that is the way it is now for geospatial information.

With a recent federal grant, the Harvard Center for Geographic Analysis has begun to build HHypermap (Harvard Hypermap), a system for building, maintaining, and deploying a public registry of web map services of various types, starting with OGC and Esri services. A preliminary investigation suggests there are millions of map services spread across hundreds of thousands of servers around the world, and this number is growing.

Short video demo https://youtu.be/lbSCw6LHsWg.

The platform we are developing discovers servers, harvests metadata from them, organizes and indexes the metadata, makes them searchable by time, space, keyword, and other characteristics via a public API. A specialized client allows users to interactively navigate and visualize the geographic distribution of map layers even when the result of a search returns millions of datasets. Layers discovered can be used within WorldMap (http://worldmap.harvard.edu) or within other systems.

Map services visualized within in WorldMap are reprojected to a common space and cached to minimize load on remote servers. Uptime statistics and usage statistics are continually gathered for all servers and layers, information which is incorporated into search result weightings.

While the system we will demonstrate runs within WorldMap, any mapping client will be capable of searching the API and returning layers to load to their own system.

The presentation will include a description of the project and where it is headed, as well as an overview of the system architecture and a live demonstration.

Slides

Session details
Speaker(s): Session Type: Experience level:
Beginner
Track: Tags:
ogc
Schedule info
Session Time Slot(s):
Tuesday, May 3, 2016 - 17:40 to 17:55

Comments