Other Case Studies
The SMART partnership supports improved law enforcement and biological monitoring in protected areas around the world. They came to Refractions to build a software tool to help meet their needs.
Biodiversity BC and its government and non-government partners were looking for regional-level environmental statistics. Refractions proposed a new approach to generating GIS summaries, using the power of a relational database and web tools to provide GIS analysis to users who previously had no access to it.
The City of Vancouver collects water readings from thousands of residences and businesses, but has never optimized the pedestrian routes the readers take. Refractions developed algorithms to create the most efficient routes of the correct length for readers.
Refractions provides full-service support for the British Columbia Digital Roads Atlas – systems design, maintenance, data conflation, client service, and rapid response.
Refractions converted a legacy habitat modelling system based on ArcView 3.X to ArcGIS 9.2, and automated the workflow to provide faster turnaround time for model runs.
The British Columbia Ministry of Forests needed a tool to conflate multiple roads databases into a single working layer. Refractions delivered the algorithms and a user interface based on the uDig platform.
Refractions developed an ArcMap extension to automate the calculation of standard reports joining a massive shape-file archive with a large Oracle database.
UN FAO needed a data collection tool that could run disconnected and didn't have a per-seat licensing cost. Refractions delivered a simple tool using the uDig desktop platform.
The Open Geospatial Consortium runs regular “testbed” projects to field-test new concepts in geospatial interoperability. Refractions was a part of the OWS-3 initiative, and built a uDig-based “GeoDSS” client to provide access to several other OGC standard services, including a prototype GeoVideo service.
Rento is a free web service that makes it easy to find a place to rent in Greece. Powered by PostgreSQL and PostGIS, it provides ad listings with photographs displayed on a map. Searching is accomplished through map navigation and natural language processing, with the search engine being capable of answering complex (spatial) questions such as "flat near the University of Athens" or "loft up to 800 euros near a metro station" (in greek).
Open Source Web Mapping Pilot – AT&T Wireless
AT&T Wireless (now Cingular Wireless)
AT&T Wireless had built an internal mapping application to drive a retail store locator and wireless coverage map. Using a proprietary software stack (MapInfo MapXtreme on Oracle Spatial) the application was deployed internally to a small group of clients (the inside sales team).
Despite minor complaints about speed, the original application served the internal clients adequately enough. However, when the time came to roll the services out to the public web site, AT&T had a major problem – there was no way to scale-up the application to service the loads that the public site would generate.
- The current site was already running noticeably slowly with just a few hundred users, on expensive Sun hardware. Substantially more hardware would be required to support the public site.
- The licensing cost of the MapXtreme/Oracle Spatial combination would dwarf the hardware cost as the amount of servers was scaled up. Total costs just to support such a mapping application would be into seven figures
AT&T needed a new solution that could scale to handle the higher load, without breaking the bank.
The goal was to show that an open source solution could perform exactly the same tasks as the existing software stack – to replicate the existing application with a different backend.
AT&T had short deadlines, and Refractions worked with their team to implement a complete system within their timeframe. Refractions installed the software, prepared road network data and loaded it into the database, created sophisticated map styles to and cartography, and wrote a new geocoding module for PostGIS to handle address look-ups.
The entire proof-of-concept, from an initial meeting to a fully operational system, was delivered in under two months.
Immediately, AT&T's new open source application ran several times faster than their original system, even though it was deployed on commodity Intel hardware.
Additionally, because there was no software license cost AT&T had the option of scaling their application for only the cost of hardware. Finally, because the new application ran faster, on commodity hardware, the hardware scaling cost was reduced as well.