FME Users Tackle Really Interesting Problems: Messy Addresses, Found Feet, Open Data Portals
Last week I attened the FME User Conference in Vancouver, BC. The user presentations made up a good deal of the event. These are the ones that I found most interesting, memorable or just cool.
The “New” 911 Geocoding
With the increased interest in many households having just a cell phone and not a landline, there are new challenges for emergency notifications. An unnamed city wanted to be able to notify cell phone users who lived in its boundary. California CAD Solutions, a long time GIS and CAD reseller and consulting firm, thus had to clean the address data of people who subscribe to the two local cell providers, Verizon and AT&T. The address lists are created from the address subscribers given when they sign up for service or later moved.
The data was a mess and originally geocoded with only 65% accuracy. Using a number FME workspaces, including one the company calls MC Hammer, used to fix an issue with MC and Mc type addresses, the data were cleaned. Those routines are run regularly as new data comes in with new subscribers or changes to existing ones.
California CAD’s work brought the match rate to 99.6%. So, now if there is a spill, for example, the city can run a buffer and notify cell phones owned by people who live within the area. For now the solution does not use a Web service, but California CAD suggests that a connection to something like Twilio, a Web service for notification, might be the next step.
The highest praise for presenter Amanda Graf of California CAD? The first question from the Esri geocoding product manager: “Can I have your card?”
FME Meets the BC Coroners Service
Randy Dalit works for the BC Coroners Service. More specifically, the Identification and Disaster Response Unit. That group is responsible for determining which bones belong to whom and how they died. Some cases are current and others are rather old.
Dalit described how the service uses FME to connect its various databases and spatialize them. The Enhanced Identification GIS integrates spatial, temporal, biological and other forensic descriptors used to compare missing persons against unidentified bodies. Most valuable features, it seems, are the ability to search on other cases near new ones and the ability to find cases with non-spatial, connections.
The story he told that piqued my interest related to a number of feet found in the Salish Sea since 2007. It was big news up in British Columbia, but didn’t make it to me in Boston. Hikers and fisherman kept finding feet, typically in sneakers, along the shore from British Columbia down to Washington State There were left feet and right feet, feet in New Balance shoes and feet in Wal-Mart ones. The brand of shoe and its ID number helped determine how long the foot might have been floating before it was discovered: It could not have been before the shoe was available for sale!
A variety of tools including DNA helped the service determine how the feet got to their final destinations. Some, per Dalit, where the feet of people who had jumped off of bridges to commit suicide. Their bodies decomposed a bit and eventually the feet in their very buoyant foam shoes rose to the surface and floated downstream. This was one of the coolest uses of FME and GIS I’ve ever heard.
And, for the stats folks: Since the implementation of FME and the integrated coroner’s database the Identification and Disaster Response Unit case load has dropped 6%. In short, Dalit said, “FME saves me time.” You can find Canada’s Missing database (including some shoes) online. That’s where the public can help Dalit’s team tackle unsolved cases.
FME Powers CKAN Open Data
I’ve written about Mark Loudon before. He’s the fellow behind Save the Rain, an app built for a Canadian climate change contest. These days, however, he’s working on an open data portal for the City of Surrey, BC, home of Safe Software. Loudon, part of a team of 15 city GIS staffers, is linking CKAN, an open source data portal, with FME.
Surrey has two main GIS resources for its citizens. One is an interactive map called COSMOS built on ArcGIS Server. The other, currently in development, will be the open data portal.
Surrey wants consistency in its mapping apps. In particular, Loudon explained, a key design goal is ensuring open data previewed on the CKAN site look just as they do in COSMOS. COSMOS renders its map data served from an Esri Map Service via .lyr files. Unfortunately, the .lyr files typically include multiple features, while individuals downloading data may only want to visualize, and then download, just one type of feature. Loudon used FME to build a complex looking workspace that extracts the definition of each feature class (and its symbology) from the .lyr file and uses it to populate the map in the CKAN app. The powerful FME workspace means the same Esri Map Service supports both COSMOS and the open data site.
My favorite part of Loudon’s presentation was afterward, when several members of the audience gathered around the speaker to ask follow-up questions. There’d been a question in the formal Q&A about if the Esri-based metadata was available for download with the data. Loudon had said basically, “we are working on it.” No worries, a fellow from York, BC told him. His group had already figured a workflow for just that and shared his card. A woman from Calgary was anxious to do just want Surrey was doing with open data. She was jazzed that all Surrey’s code was on GitHub and took Loudon’s card with a big smile on her face. I enjoyed watching “how GIS happens” and confirmed yet again that the tech is just a small part of GIS, it’s about the people that truly power the solutions.
Disclosure: Safe Software covered my travel and lodging for this event.