blog authors
past blog entries

Welcome to the Kellylab blog

geospatial matters

Please read the UC Berkeley Computer Use Policy. Only members can post comments on this blog.

Entries in open source (55)

Tuesday
Nov222005

Open Source GIS Review (plus dynamic web modeling)

Aaron Racicot up at Ecotrust has recently posted his review of open source GIS software that he presented at the Oregon State University GIS Day. It's great material, providing an excellent overview of what packages are out there, what they're good for, and what they're not. He's also made some excellent posts on the OPENNR list about his work integrating open source tools in a web-based decision support tools for natural resource management.

Monday
Oct102005

Geodata Provider for the GIIF

Finding geodata can be frustrating. There is no Google for publicly available geodata, and even within our own lab data exists almost exclusively on individual workstations, fallow and unsearchable, like oil tragically buried under a wildlife refuge. However, the GIIF has the funding, resources, and collective brainpower to alleviate this problem (uh, with the geodata, not the oil). Behold, my vision of the future: a centralized geodata provision service (or G-P-S, not to be confused with Gap, Inc. stock). I see this as a centralized repository for data relevant to us, stored in a standardized, possibly version-controlled manner that enforces metadata creation, and searchable through a usable web interface that allows queries on metadata and spatial queries (i.e. show me all the transportaiton data in Alameda County created after 2003). Here are my thoughts. Does anyone else think this is a good idea or should I seek therapy? What other features would you like to see? Does this software already exist? Please comment!

Geodata Server

Who it would serve

  • GIIF users
  • CNR
  • UC Berkeley
  • general public
  • desktop users
  • GIS application developers (including webGIS projects)

Note this all implies some kind of access control

What it would serve

  • our own data (VTM, SOD models, fire models, wetland maps, etc.)
  • external data (TIGER, FRAP, CaSIL, US Census, anything that allows redistribution
  • spatially indexed links to external data where redistribution is prohibited
  • symbologies (recommended symbologies and/or visualizations for complex data, i.e. Mapserver layer definitions and ESRI .lyr files)

Geodata Portal Application

How it would serve data

How it would store data

  • extents stored in a geodatabase for retrieval via spatial query (box selection, search by distance, etc)
  • metadata stored in a database for retrieval by metadata query (by date, by source, by type, etc)
  • actual data maintained in original format (versioning? how do we deal with data in external geodatabases?)

How users would interact with the data

  • search for data via spatial and non-spatial queries (see above)
  • automatically generated previews (see what you're getting)
  • AJAX development techniques used whenever appropriate
  • submission: users should be able to submit new data, or modified versions of existing data
  • enforced metadata: all data subject must have certain kinds of metadata before being accepted (date created, source, accuracy. We can derive a lot automatically, like extent, file type, projection, file size, etc)

Other ideas

Possible models/inspirations

Possible software tools/components

Note extreme open-source bias... Definitely open to other suggestions.

Tuesday
Oct042005

GRASS and QGIS preview

Apparently GRASS and QGIS integration is coming along nicely: check out this really cool movie showing some of the QGIS/GRASS features in development. For those who don't know, GRASS is an open-source GIS software package that has been around forever and is fairly powerful, but has remained somewhat inaccessible by its steep learning curve and lack of a usable GUI. QGIS is another open-source GIS package with a nice ArcView-like GUI but lacking in deep functionality. Together, they fight crime

.

Friday
Sep162005

Dynamic GRASS map and PHP

Here's a small demo of PHP displaying a dynamically rendered GRASS map. The script just makes GRASS shell commands, and GRASS renders a PNG to a web-readable directory. I don't really understand the GRASS stuff, but the PHP looks dead simple. Tim, if the GRASS commands look relatively simple to you, maybe we could try this with a simple SOD model some time in the unspecified future.

Wednesday
Sep142005

FOSS development brainstorming

There's an interesting discussion going on over on the OPENNR list regarding development of an open source image analysis package, in case anyone's feeling particularly nerdy today.

Page 1 ... 2 3 4 5 6