Spatial Statistics on the Geospatial Web

Abstract

The Geospatial Web comprises standards and software for a wide variety of applications, ranging from typical GIS (geographic information systems) functionality, via emergency management and monitoring use cases to distributed workflows [27]. Functions include data access and visualization, but also workflows are published as web-accessible resources. Web processing is a well-known concept for distributing processing capabilities on the web, and the OGC Web Processing Service (WPS) [24] standard defines a service interface for this [28]. Spatial statistics is a branch of statistics that focuses on choosing and fitting models and predicting observables from spatial data, and focuses on cases where the spatial locations matter. Spatial statistics adopts different types or representations of spatial and spatio-temporal data, for example points, lines, polygons, grids and trajectories [17]. A substantial part of the applied spatial statistics community works with R [2,21], a language and environment for statistical computing [19]. Other GIS applications use tools for spatial data analysis written in Python scripting language [25]. This means that complete analyses can be communicated, and reproduced, by scripts, along with input data. Web processes in the Geospatial Web can use a wide variety of input and output data in different data models and encodings (such as GML) and formats (such as NetCDF, XML, but also service interfaces such as WFS). How a processing backend actually carries out the analysis is typically invisible to the user. The motivation for web-based processing comes, on the one hand, from scarcity, for example of computational power, bandwidth or knowledge. Mainstream technologic developments such as grid and cloud computing provide an infrastructure to outsource processes to high performance environments and facilitate the integration of heterogeneous resources [9] and real-time data handling. On the other hand, publishing processes implies sharing capabilities without dependencies on the scripting environment and allows for an open scientific discourse that includes researchers who are not domain-experts or experienced programmers in a particular language [19]. With this work, we investigate how scripted geoprocesses can be integrated into a standardized web service interface. As a solution we describe a generic annotation concept. We apply it to the example of R scripts published in an OGC WPS. It may later on be applied to other scripting languages (e.g. Python) and server environments as well. R provides several packages for web service interactions namely Rserve [26], websockets [12], Shiny [23], RWebServices [14] and RevoDeployR [22]. They execute either code blocks …

Showing 1-10 of 25 references

GDAL -Geospatial Data Abstraction Library, Version 1.9.2

  • 2013

How to Write Doc Comments for the Javadoc Tool

  • Oracle Corporation
  • 2013
1 Excerpt

Quantum GIS Geographic Information System

  • Gis Quantum, Team
  • 2013

R: A Language and Environment for Statistical Computing

  • R Team
  • 2013
1 Excerpt

DCMI Usage Board

  • 2012

The R software environment in reproducible geoscientific research

  • Edzer Pebesma, Daniel Nüst, Roger Bivand
  • 2012
1 Excerpt