Volume IV-4/W2
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-4/W2, 31-35, 2017
https://doi.org/10.5194/isprs-annals-IV-4-W2-31-2017
© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 4.0 License.
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., IV-4/W2, 31-35, 2017
https://doi.org/10.5194/isprs-annals-IV-4-W2-31-2017
© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 4.0 License.

  19 Oct 2017

19 Oct 2017

DETECTION OF BEHAVIOR PATTERNS OF INTEREST USING BIG DATA WHICH HAVE SPATIAL AND TEMPORAL ATTRIBUTES

R. W. La Valley1, A. Usher2, and A. Cook3 R. W. La Valley et al.
  • 1OGSystems, Inc., Data Scientist and Senior Statistician, 14291 Park Meadow Dr # 100, Chantilly, VA 20151, USA
  • 2Digital Globe Intelligence Solutions, Chief Technical Officer, 4350 Fairfax Dr., Ste 950, Arlington, VA 22203, USA
  • 3Digital Globe Intelligence Solutions, 4350 Fairfax Dr. Ste. 950, Arlington, VA 22203, USA

Keywords: Geospatial, Temporal, Aggregation, Location, Z-Curve, Space-Time Boxes, Geo-Temporal Hashing, Big Data

Abstract. New innovative analytical techniques are emerging to extract patterns in Big Data which have temporal and geospatial attributes. These techniques are required to find patterns of interest in challenging circumstances when geospatial datasets have millions or billions of records and imprecision exists around the exact latitude and longitude of the data. Furthermore, the usual temporal vector approach of years, months, days, hours, minutes and seconds often are computationally expensive and in many cases do not allow the user control of precision necessary to find patterns of interest.

Geohashing is a single variable ASCII string representation of two-dimensional geometric coordinates. Time hashing is a similar ASCII representation which combines the temporal aspects of date and time of the data into a one dimensional set of data attributes. Both methods utilize Z-order curves which map multidimensional data into single dimensions while preserving locality of the data records. This paper explores the use of a combination of both geohashing and time hashing that is known as “geo-temporal” hashing or “space-time” boxes. This technique provides a foundation for reducing the data into bins that can yield new methods for pattern discovery and detection in Big Data.