GeoWorld

Let's Talk Sensors WHERE IT'S ABOUT TIME I BY ERIK SHEPARD n my last column, I discussed events as a distinctive type of spatiotemporal data (see "It's Not a Party, It's an Event," GeoWorld, August 2011, page 30). Generally, events are thought of as discrete activities or occurrences that happen at specific points in time and usually at specific locations (e.g., car accidents, tornados, birthday parties, etc.). The virtual explosion of sensor data, from cellular to smart-infrastructure devices (think SCADA or Smart Grid), has provided a wealth of new sources for geospatial data. Much fieldwork to date has focused on building core capabilities and extensive basemap data, but acquiring new data has been a time- and cost- intensive process (e.g., aerial photography, remotely sensed satellite data, field surveys, etc.). Much of the late 1990s and 2000s were spent in data acquisition. Sensors Provide New Opportunities The advent and widespread deployment of sensors shifted the paradigm. Rather than data collection being built on expensive, extensive and infrequent activities, data collection can be accomplished for some types from (relatively) inexpensive devices deployed across a wide area, leveraged so the whole definitely is greater than the sum of its parts. Event (or discrete) data typically are fixed in space and time, with varying attribution (i.e., something happened), while change data typically are fixed in space but vary in attribution through time. Dynamic data typically vary in location over time, but are fixed in attribution. Sensor data represent a fourth type, typically Erik Shepard is principal of Waterbridge Consulting; e-mail: erik@ waterbridge.biz. 30 providing a measurement at a particular place (or places) in space and time (i.e., varying attribution through space and time). Interestingly, some practi- tioners classify such data as "stationary" in keeping with the fact that they typically were associated with fixed locations such as weather stations or telemetry points. But smartphones and other cellular devices have provided opportunities for sensors that aren't necessarily stationary. GEO W ORLD / N O VEMBE R 2O11 Sensors Enable a Geospatial Renaissance GIS has become a mainstream technology and fol- lowed a path similar to the typical innovation s-curve: emergence, rapid development, mainstream accep- tance and, finally, maturity. The Geospatial Information and Technology Association (GITA) organization and conference is an excellent case study; the conference began in the late 1980s and was heavily attended through the 1990s and early 2000s. Attendance began declining a few years ago, and GITA has struggled with its identity during the last few years. In 2011, GITA reorganized and ended its main conference. Geospatial professionals are well positioned as "subject- matter experts" for such spatiotemporal data. But as anyone familiar with Paul Harvey knows, the story doesn't end there. GITA signed a memorandum of agreement with Utilimetrics, the utility organization that, until recently, focused on automated meter reading and infrastructure, but now focuses on utility automation in general. The Utilimetrics conference, Autovation, offered a geospatial track in conjunction with GITA in September 2011, and, by all appear- ances, the partnership looks to be a promising one. But what's underlying this partnership? Certainly geospatial has been heavily leveraged by the utility industry and is a core requirement for next-generation programs such as smart metering and the smart grid. But the real crux is that sensor data are providing a new opportunity for renaissance in the geotechnology field. Geotechnology is key for leveraging, managing, visualizing and analyzing spatiotemporal sensor data. And geospatial professionals are well positioned as "subject-matter experts" for such spatiotemporal data. The term "sensor web" was introduced in the late 1990s to describe a distributed sensor network as a geospatial source—but being more than the sum of the individual sensors. Rather, it's a collective with analytical functions to aggregate data, find patterns and filter outliers. The Open Geospatial Consortium (OGC) provides the Sensor Web Enablement framework to coordinate heterogeneous sensors. SensorML is an OGC stan- dard that leverages GML, which, in turn, contains spatial and temporal components. Sensors (and sensor data) are spatiotemporal at their core. Government Special Issue

• Cover