It is all good to have access to large volume of data generated with high velocity which normally spans variety of domains and usually comes with levels of veracity. However, the importance of data becomes visible only when we turn it into a value. In particular, in the domain of “Big Spatial Data”, we deal with massive amounts of real-time spatial and spatio-temporal data obtained from billions of sensors, location-aware devices, remoting sensing satellites, and various models of the physical world. The use of Big Spatial Data spans a variety of applications including social networks, earth sciences, transportation, communication networks, online maps, smart cities and urban planning, remote sensing, and crisis and evacuation management, to name but a few. Turning Big Spatial Data into value is challenging and requires introduction of fundamentally new spatio-temporal algorithms, methods, and systems that can process, mine, and analyze massive amounts of fast and heterogeneous spatio-temporal data in a timely manner.

Keynote 1

Abstract: The world of digital data discovery has been revolutionized by the ability to index and rapidly search information on the web, on social networks, and in business transactions (E.g. 45 billion web pages have been indexed allowing to search and discover them in ~ 0.5 second). Despite all these advances, insights from geospatiotemporal data -- information collected across space and time, such as weather data and models, satellite and aerial imagery, map information and data from devices and sensors from the Internet of Things (IoT) -- remain relatively “dark” and are difficult search and discover due to its complexity, formats but most importantly to its enormous size. By some estimations, geospatiotemporal data is larger than any other data including social data growing by many hundreds of Exabytes per month. In this presentation we discuss new approaches for scalable geospatiotemporal data processing and analytics. The talk will include a series of examples illustrating complex analytics on multi-modal data sets including applications to weather, agriculture, and renewable energy and land use recognition.

Bio: Dr. Hendrik F. Hamann is currently a Senior Manager and Distinguished Research Staff Member at the IBM T.J. Watson Research Center, Yorktown Heights, NY. He received his PhD from the University of Göttingen in Germany. In 1999 he joined the IBM T.J. Watson Research Center, where is leading the Physical Analytics and cognitive Internet of Things program. Hamann’s current research interest includes sensor networks, sensor-based physical modeling, machine-learning, artificial intelligence as well as big data technologies. Hamann has authored and co-authored more than 100 peer-reviewed scientific papers and holds over 110 patents and has over 100 pending patent applications. Dr. Hamann is an IBM Master Inventor, a member to the IBM Academy of Technology and has served on governmental committees such as the National Academy of Sciences, the National Science Foundation and as an industrial advisor to Universities. He won several awards including the 2016 AIP Prize for Industrial Applications of Physics. He is a member of the American Physical Society (APS), Optical Society of America (OSA), The Institute of Electrical and Electronics Engineers (IEEE) and the NY Academy of Sciences.

Keynote 2

Abstract: Much in our daily lives is limited by time. There are 86,400 seconds in a day and we are in a constant battle to make the most of them. More time spent on research may mean missing out on time with our families or losing sleep. Those of us in the “Big Spatial Data” realm are far from immune to this struggle. The greater the volume of data the more time it takes to process that data and with the truly massive amounts of information that are available today, that loss of time may become untenable. One of the most efficient solutions to this problem is to use a distributed computing framework to parallelize the work. However, most of these frameworks utilize a single dimensional key-value store which is a problem when you need to preserve locality in multiple dimensions. This is where the GeoWave open source framework comes in (https://github.com/locationtech/geowave). At its core, GeoWave is a software library that connects the scalability of various distributed computing frameworks and key-value stores with modern geospatial software to store, retrieve, and analyze massive geospatial datasets. The multidimensional indexing algorithms in GeoWave preserve locality in any number of dimensions which decreases your processing time by multiple orders of magnitude. Jobs that previously took hours can be performed in seconds which makes it possible to fully utilize these massive datasets and makes that battle against the clock much easier for analysts and researchers.

Bio: Michael Whitby is a Software Engineer in the Radiant Solutions group of Maxar Technologies (formally Digital Globe). He works on the GeoWave open source project and has had his writing about the framework published in Advances in Spatial and Temporal Databases.
He is a U.S. Navy veteran and received his masters in Software Engineering from Pennsylvania State University.