Difference between revisions of "Case Studies"

From OSGeo
Jump to navigation Jump to search
Line 53: Line 53:
  
 
Todd Buchanan's master thesis compared ArcGIS 9.0 and GRASS 6.0 for a use case: "Characterization of Urban Sprawl for Eugene-Springfield, Oregon". He took detailed notes using both and found "ArcGIS is more straightforward than GRASS. However, GRASS proved itself fully capable of performing the operations required by the case study and outperformed ArcGIS in several categories."
 
Todd Buchanan's master thesis compared ArcGIS 9.0 and GRASS 6.0 for a use case: "Characterization of Urban Sprawl for Eugene-Springfield, Oregon". He took detailed notes using both and found "ArcGIS is more straightforward than GRASS. However, GRASS proved itself fully capable of performing the operations required by the case study and outperformed ArcGIS in several categories."
 +
 +
Note: GRASS is even better now and has been integrated with the user friendly qgis since this study.
  
 
=== uDig Case Studies ===
 
=== uDig Case Studies ===

Revision as of 23:12, 25 April 2008

This page lists Case Studies of projects that include Geospatial Open Source components.

Case Studies

Web Clients

Desktop Clients

Review of Open Source Desktop Clients

April 2008 http://www.spatialserver.net/osgis/

A comprehensive feature by feature comparison of all the Open Source desktop clients.


  • Municipality of Trento (Comune di Trento), Italy, migration (2005-today): especially improvements in GRASS GIS are financed: new digitizer tool and better high quality cartographic output.

Email threads discussing merits and disadvantages of OSGeo Desktop

18 April 2008: Geowanking: ArcView 3.1 replacement

"[Oxford Archaeology] are looking at gvsig and qgis as the main options [for ArcGIS], gvsig because it can use cad data, and qgis because we like the grass integration and it's slightly more user-friendly for english speakers (the translated version of gvsig still has some spanish bits in it). With slight changes to our work-flows, we are finding that these two packages will do almost everything we need a gis to do, with the exception of producing high-quality illustrations. To achieve this we are currently looking at export to svg or postscript for final editing in inkscape, but that's a work in progress."

24 April 2008: OSGeo-Discuss: Can I do the same GIS tasks with OS (as with ESRI)?

This email thread provides numerous opinions for and against using Open Source Geospatial Software.

Paul Ramsey notes there is not a direct Open Source replacement for Arc View:

I think it's contingent on us as evangelizers to not over-sell. I would not recommend QGIS or any other open source desktop to someone whose prior experience was Arc* until I had a clear understanding of the use case. In response to the query "can I replace ArcView with open source", my answer is "in general, no, but maybe for a specific use case".

John Callahan provides a good summary:

I've been an ESRI user (AV 3.x, ArcGIS 8/9, ArcIMS, ArcGIS Server, ArcSDE) for 12+ years and have recently started exploring FOSS software. And I haven't disagreed with any of the responses so far. You will definitely need multiple programs to do what a single ESRI program can do. IMO, this is a good thing. One of the main reasons for my migration is I'm tired of running large, complicated, expensive software and all the extra baggage that comes with it to use only 10% of what the software can do. (see ArcGIS Server.)

You can do all the analysis and more of ArcGIS Desktop and extensions using GRASS, QGIS, SAGA, GeoTools, GDAL/OGR, PROJ4, or R Statistics along with a programming language like Python, Java or others. (IMO, this is a better solution than ESRI.) You can do just about anything you want on the web server end with MapServer, GeoServer, FeatureServer (and maybe TileCache or GeoNetwork for metadata) with any of a dozen or more clients (OpenLayers, ka-Map, MapGuide etc...). And you can do a lot of database work with Postgres/PostGIS, a much simpler, less costly solution than ArcSDE+RDBMS. And I wouldn't count out the role of free, non-open source packages like Google products and Oracle Express (11g should have Spatial included.)

From my experience (limited in the FOSS world), I have found three basic hurdles:

  1. Cartography. Whether on the screen, PDF outputs, or print publications, ArcMap is easy and looks great. (Although R Statistics produces better looking charts and graphs than ArcGIS.)
  2. Versioned editing. This is important for groups with multiple concurrent editors or that has a particular hierarchical workflow with their GIS data.
  3. Storage and serving of very large (50+ GB) raster datasets. PostGIS does not support rasters yet; Oracle Spatial does though. I'm still not sure if storing rasters in a database is a good idea but ArcSDE sure makes it easy, and with good performance when used in conjunction with other ESRI products.

Current work addressing Open Source holes include:

GRASS

August 2005, Comparison of ARCGIS 9.0 and GRASS 6.0

Todd Buchanan's master thesis compared ArcGIS 9.0 and GRASS 6.0 for a use case: "Characterization of Urban Sprawl for Eugene-Springfield, Oregon". He took detailed notes using both and found "ArcGIS is more straightforward than GRASS. However, GRASS proved itself fully capable of performing the operations required by the case study and outperformed ArcGIS in several categories."

Note: GRASS is even better now and has been integrated with the user friendly qgis since this study.

uDig Case Studies

http://udig.refractions.net/gallery/ Collection of uDig case studies put together by Refractions.

Web Services

Army Corps of Engineers Wetlands Regulatory program

25 September 2007. Powerpoint presentation from FOSS4G Conference

The Army Corps successfully integrated a Proprietary Oracle Database, ARC SDE and Analysis tools, with Open Source WMS and WFS services, and free viewers like Google Earth and Google Maps.

Databases

PostGIS Case Studies

A collection of PostGIS Case Studies collected by Refractions.

Complete Infrastructure

Valencian Regional Ministry of Infrastructure and Transport

in Spanish (view English translation by Google)

Valencian Regional Ministry of Infrastructure and Transport is under a project of migrating "all" systems to open-source software. As a matter of fact, that was the reason to build gvSIG. All geospatial infrastructure (previously with ESRI) has been moved to open source (gvSIG, PostGIS, MapServer, deegree, GeoNetwork opensource).

Hydrographic Confederation of Guadalquivir River, Spain

in Spanish (view English translation by Google)

A migration of a big part of ESRI components to open-source has been made, with use of gvSIG, GeoNetwork opensource, MapServer, GeoServer, deegree.

ROI from Geospatial Open Standards, study by NASA

April 2005

National Aeronautics and Space Administration Geospatial Interoperability Office studied the Return On Investment from using Geospatial Open Standards. The study showed "There is a significant improvement in functionality and mitigation of cost when using open as opposed to proprietary standards. The project that implemented geospatial interoperability standards had a risk-adjusted ROI, or "Savings to Investment" ratio, of 119.0 percent throughout the five-year project life cycle."

Predicted ROI for Australian Government to investment in a SDI

October 2007

ACIL Tasman study reports a 10 to 20 times ROI should the Australian government invest in a Spatial Data Infrastructure.

Open Source Policies

These documents provide compelling reasons for selecting Open Source and Open Standards.

Australia

April 2005, [Guide to Open Source Software for Australian Government Agencies]

It can offer original solutions to problems not addressed by proprietary software and it has the potential to lead to significant savings in Government expenditure on information and communications technology (ICT).

New Zealand

13 March 2008, [Ministry of Justice Open Source Strategy Paper]

Proposes when Open Source should be used over Proprietary software, and concludes that Open Source better than Proprietary, all other things being equal.

US Navy

April 2006, [US Navy Open Technology Development Roadmap Plan].

Describes how the US Navy will migrate to Open Technology Development when purchasing and building software.

Open Technology Development combines salient advances in the following areas:
* Open Standards and Interfaces
* Open Source Software and Designs
* Collaborative/Distributive culture and the and online support tools
* Technological Agility

Also, a much shorter news article on the same topic from March 2008. [The Navy will acquire only systems based on open technologies and standards].

Why write a Case Study?

A case study discusses how a specific situation was identified, which solutions were investigated and selected, and a summary of the results.

Solution architects use relevant case studies to support a solution they are building.

Writing a case study about your project gives others the confidence to follow in your footsteps which in turn increases investment in your tools of choice. Publishing your experience gives positive feedback to those involved in your project, reinforcing your success, encouraging colleagues to take on additional challenges, allowing the project to continue to grow.

We would like to see case studies which cover a range of use cases, especially from organizations with a low risk tolerance. Many organizations approach Open Source in little steps gradually integrating it with existing infrastructure and find targeted small case studies valuable.

What to write?

One to three pages is good, with a graphic per page.

The case study should cover:

  • Problem
  • Evaluations
  • Implementation, including integration with other infrastructure
  • Problems faced and how they were overcome
  • Return on Investment
  • Future plans
  • Who else might benefit from your experience

More hints found by googling: How to write a Case Study

How to publish?

Consider submitting an article to the OSGeo Journal. Here are the basic guidelines for authors.