[postgis-users] PostGIS Usage Survey

[Carl Anderson]

I wear two hats

Georgia Speleological Survey
   part of Web app to record and render points of interest
     used for trigger based projection of data.
   used for map generation

  Web app is PHP/Mapserver based
  Mapping tools are GRASS based

-----------------------------------------------------------------------

Fulton County, Georgia USA

   web based entering and editing of,
   web based vizualization of complex table joins of,
   management and generation of time thematic data versions of,
   proximity analysis using,
   extraction of complex SQL queries of (as shapefiles),
   extraction of production maps of,

     Cadastral data (280,000 Parcels)
     Zoning data
     LandUse (Future) plans
     Annexation cases
     Sanitary Sewer AM/FM
     Centerline (new roads as part of permitting processing)
     Addresses (new addresses as part of permitting process)
     Building Permits
     US 2000 Census SF1 files (visual reporting only)
     US 2000 Census SF3 files (visual reporting only)
     Specimen tree inventory
     GPS monumentation inventory
     Sewer Flow Monitoring (mostly RDBMS info)  
   Automated Address Geocoding
     Management of data for use in GRASS
   (mostly 'cause GRASS is horrible at large dataset management)
     DTM (36 million data points)
     Elevation Countour (88 million vertices)
     Spot Elevations (50 thousand points)
   used Post GIS to assemble and reselect data based on proximity and
    validity while generating a DEM from the three data sources.

I know that this scale of data can easily be exceeded. my point in 
showing the scale is that PostGIS can manage datasets at least this 
large (with little difficulty).

PostGIS data is used with
   Arc/Info  (custom scripts)
   Arcview Unix & PC (custom scripts)
   Mapguide (currently write only, OLE Db limitation)
   GRASS-5
   Mapserver
   Perl (custom scripts to extract/import shapefiles)
   PHP (pass through for object insertion/editing)

The advantages of PostGIS on PostgreSQL: (IMHO)
   easy to create trigger based data consistency rules
   shared, concurrent data editing
   web based concurrent data editing
   easy to replicate data to query-only servers
   relatively lightweight RDBMS server footprint   leveraging other 
Open Source projects results in quickly expanding features
   we can share project burdens with others without licensing issues
   we can establis development machine w/O having to justify the cost


A bit of a history:
   In 1998 and 1999 I hacked Shapelib to add limited geospatial support 
into   BLOBS for Oracle (super ugly code never made public).  For 
various reasons I
   was running PostgreSQL at the same time.  When PostGIS came to my 
attention
   I moved the hacks over to PostGIS and modified the custom scripts to 
support
   postgres.


I will package up some of the better custom scripts (finally).


-- 
----------------------------------
Carl Anderson
candrsn at mindspring.com

New Hampshire tells us "common sense for all"
     where can I get mine?
----------------------------------