Scaling Up Spatiotemporal Analytics with GeoMesa

Storing and analyzing large spatiotemporal datasets is becoming a necessity as geo-tagged data becomes increasingly ubiquitous. GeoMesa (http://geomesa.github.io/) is an open-­source, distributed, spatio-­temporal database built on top of Apache Hadoop and Apache Accumulo. GeoMesa is capable of ingesting high­-rate spatio-temporal data such as Twitter, FAA flight tracks, and near-­realtime geolocated sensor data. It also leverages Storm and Spark to provide streaming data analytics.

Hunter and Andrew will provide an overview of how to adapt existing open-source geospatial libraries to build an open-source analytic pipeline for large geospatial datasets. They’ll discuss GeoMesa architecture, query-­planning, and OGC mapping capabilities along with demos of analytics including interpolated space-­time queries, track similarity rankings, and distributed K­-Nearest Neighbors.

Biographies

Andrew Hulbert is a software engineer at CCRi specializing in distributed indexing and streaming analytics. He is an Eclipse Foundation committer on the LocationTech/GeoMesa project and has been using Accumulo for three years. Prior to joining CCRi, Andrew worked on several projects for the Department of Defense providing large­-scale data management, storage, and indexing using Hadoop as well as traditional data­-warehousing in parallel RDBMS environments. Andrew spends his spare time hiking, playing tennis, and enjoying the local breweries around Charlottesville. He earned his B.S. in Computer Science from the University of North Carolina at Chapel Hill.

Hunter Provyn is also a software engineer at CCRi with a background in geospatial and statistical analysis. He has significant experience developing backend software for geospatial analytical tools in Java and Scala. Mr. Provyn also has experience with the Hadoop Map/Reduce framework, Accumulo distributed database, Akka actors, and Kafka-Storm topologies. Previously, he worked at BBN Technologies as a Systems Engineer and Data Analyst, where he worked on numerous network and security analytical applications and studies. Hunter enjoys cycling, beer brewing and the epicurean lifestyle. Hunter earned his B.S. in Systems and Information Engineering at the University of Virginia.

Company Profile

CCRi (Commonwealth Computer Research, Inc.) (http://www.ccri.com/) is a small analytics company in Charlottesville, VA. We design full­-stack, scalable analytics for government, military, marketing, logistics, and healthcare sectors, and implement those solutions with effective and reliable software. At CCRi, our vision is to make sense of the world by fusing data from multiple sources. We bring together a highly trained team of researchers, problem solvers, and developers, trained across a spectrum of technical fields including data scientists, systems engineers, software engineers, mathematicians, and artists just to name a few. Our culture is scientifically rigorous, data driven, and product oriented yet open­-minded and research oriented. We work to maintain an informal and creative work environment where people can think and share openly.