Presentation

Transforming Insurance, Financial Services and Telecommunications with Big Data technology

Rose Winterton, Product Director GIS and Applications, Pitney Bowes Software

Abstract

When searching for meaningful and actionable insights, accurate data quality and geospatial processing is required to make sense of the high volume, high velocity and variety of location-based data in a big data environment. In this presentation a unique approach, embedding the combination of data quality, geocoding and data conflation within big data solutions to create better business outcomes will be suggested. New insights can be uncovered when reducing millions of financial transactions or wireless call records to datasets that are aggregated by geography. Higher levels of satisfaction can be achieved when existing records and transactions are enhanced for a richer single view of the customer thus preparing the data for more advanced location analytics or visualization.

Presentation

An Overview of Databases for the Big Data Ecosystem

Keith W. Hare, JCC Consulting, Inc. and Convenor, ISO/IEC JTC1 SC32 WG3 Database Languages

Abstract

The ultimate goal of big data techniques is to be able to identify useful, usable information in a timely fashion – actionable analytics.

Refereneces

Presentation

Exploring Strategies for Optimizing Knowledge Derivation from Imagery

Dan Getman, Geospatial Big Data Solutions, Digital Globe

Abstract

Dynamically scalable compute, new storage paradigms for large raster and vector datasets, and dramatically improved machine learning algorithms have fundamentally changed how we think about architecting analysis frameworks. Balancing storage against performance and the flexibility of processing data locally with the scalability of processing data in a distributed environment has become a much more complex activity. In this discussion we will present the paths that DigitalGlobe is exploring to determine the right combination of open access algorithm development, pre-computed datasets, tile based deferred execution, map reduce driven information extraction, and event driven analysis. Through this combination of traditional and new methodologies we are pushing the boundaries of what can be done through big data analysis and knowledge extraction from raster and vector data sources. In this presentation we hope to spark discussion of the costs and benefits of these, and other, methodologies and facilitate collaborations to push those boundaries even further.

Refereneces

Presentation

Big Datacube Analytics: From Parallelization to Federation

Peter Baumann, Professor of Computer Science, Jacobs University Bremen

Abstract

Spatio-temporal datacubes are a convenient model for presenting users with a simple, consolidated view on the massive amount of data files gathered. In the EarthServer initiative, datacubes form the central paradigm for fast, versatile access to massive 3-D satellite image time series and 4-D meteoro­logical data.

Under the motto "Big Earth Data at Your Fingertips", the EarthServer initiative unites Europe, Australia, and the US in establishing a planetary scale datacube analytics federation. Among the partners are ESA, NASA, ECMWF with its 87 PB climate archive, and NCI Australia.

EarthServer successfully demonstrates the added service quality Array Databases can provide. The underlying pioneer Array Database system, rasdaman, offers an n-D array query language which forms the blueprint for forthcoming ISO Array SQL. Internally, data get partitioned and distributed transparently to the queries. Retrieval utilizes array-specific optimizations, heterogeneous hardware, and distributed processing. US magazine CIO Review has chosen rasdaman into their 100 Most Promising Big Data Solutions, being the only array engine in this lineup.

Refereneces

Presentation

Enabling access to big geospatial data with LocationTech and Apache projects

Rob Emanuele, Azavea

Abstract

This presentation will discuss tools in the open source landscape that are used to handle big geospatial data. In particular, we will focus on how Apache frameworks such as Spark and Accumulo are "geospatially enabled" by four projects: GeoTrellis, GeoWave, GeoMesa, and GeoJinni. These four projects all participate in LocationTech, a working group under the Eclipse Foundation. In particular, we will discuss how each of these LocationTech technologies implement spatial indexing (e.g. by using space filling curves) in order to provide quick access to data, and other common themes among the four projects. Attendees should walk away from this presentation understanding important parts of the Apache big data ecosystem, a set of LocationTech projects that belong to the cutting edge of enabling those Apache project's handling of geospatial data, as well as some solutions to common problems when dealing with large geospatial data.

Refereneces

Presentation

Using HPC-ABDS for Streaming Data

Geoffrey Fox, Indiana University

Abstract

We review results from two recent workshops on streaming applications and their technology. We introduce HPC-ABDs -- the High Performance Computing Enhanced Apache Big Data Stack and explain how it allows one to achieve performance of HPC and the richness and usability of Apache stack. We give some examples from robotics and data analytics. We give an initial discussion of geospatial problems from Polar science and other areas.

Refereneces

Presentation

Location empowers in a shifting big-data landscape

Dr. Jibo Sanyal, Geographic Information Science & Technology group, ORNL

Abstract

The world of location enabled big-data has heralded a fundamental shift in how we interact with our surroundings. A large number of our day-to-day activities are being driven by knowledge about how we behave, how we move, how we shop, what we care about, and how we respond in our various interactions. We are sensing the world around us at a rate that has never been done before. The data collected ranges from being unstructured to semi-structured such as geolocated tweets to highly structured such as satellite observations. Collected data has historically been used to develop inferential and predictive capabilities; however, as the volume of data has grown, the technical challenges around managing and making sense of the data has become increasingly complex. Conventional tools and techniques perform poorly in this new data paradigm. Additionally, there is a naturally accompanying notion that there are undiscovered truths hidden in the data.

This talk discusses how researchers at the Oak Ridge National Laboratory have dealt with these challenges in the shifting big-data landscape. In one application, machine learning and the use of GPU powered high performance computing, and including deep-learning, has been applied to identify human settlements all the while dealing with increasing volume and variety of satellite imagery. In another application, advanced spatio-temporal analytics enables exploration of thousands of attributes interactively. In yet another application, real-time sensor data coupled with predictive modeling illustrates the variety and velocity of operational situations. The talk touches upon key scientific and technical needs as well as makes a case for why location matters and will continue to matter in a significant way.

Presentation

Moving Features to address problems on BigData analysis

Akinori Asahara, Hitachi

Abstract

Many Bigdata applications using location data, such as traffic congestion analysis and woker managements using indoor tracks, have been rapidly increasing. Such applications require association of various geospatial data to produce more values. 'OGC Moving Features’, which is a standard targeting track-data exchange, encourages such applications. In this talk, the overview of OGC Moving Features standard and its use cases will be presented.

Refereneces

Presenter

Dr. Lea Shanley, South Big Data Innovation Hub at RENCI

Bio

Dr. Lea Shanley is a founding co-Executive Director of the South Big Data Innovation Hub at the Renaissance Computing Institute (RENCI) at the University of North Carolina-Chapel Hill. Before joining the Hub, Dr. Shanley served as a White House Presidential Innovation Fellow at NASA Headquarters, where she designed and guided open innovation and open source research strategies for planetary and Earth science. From 2013 to 2015, Dr. Shanley founded and led the Federal Crowdsourcing and Citizen Science Community of Practice, growing the community to more than 300 members from 40 agencies, and advising and leading the development of the online citizen science toolkit, which became CitizenScience.gov. From 2011-2014, Dr. Shanley was the founding director of the Washington-based Wilson Center Commons Lab, guiding strategic research in crowd-mapping, social computing, and big data, and conceptualizing and initiating the federal citizen science toolkit and projects database. Previously, she served as an American Association for the Advancement of Science/ASA-CSSA-SSSA Congressional Science Fellow and primary science advisor to Senator Bill Nelson (FL), where she made significant contributions to the NASA Authorization Act of 2010 and two other statutes. Dr. Shanley also helped to launch the Wisconsin Geographic Information Coordination Council, and spent more than 15 years conducting research and working with local, state, and tribal governments in the development of GIS-based decision support systems for city planning, environmental monitoring, coastal management, and disaster response. She holds a Ph.D. in Environment and Resources, with a focus on geographic information science and remote sensing, at the University of Wisconsin-Madison. Her work has been featured by whitehouse.gov, Fast Company, Popular Science, The Washington Post, NextGov, TechCrunch, and Vice.