Tuesday, 24 January 2012: 11:30 AM
Dynamic generation of contour images from DAP data sources using Python based web services
Room 346/347 (New Orleans Convention Center )
We have developed a web service for on-the-fly generation of contour plots of gridded fields suitable for overlay over geospatial maps in GIS applications. Most contour maps currently used for operational outlooks by meteorological services are produced by scheduled server-side applications and stored as images. With ever increasing amounts of data being made available, the related increase in required storage means static plots are not a viable solution for the delivery of all maps to end users. Contour plots are one of the most used data visualisation techniques in meteorology and oceanography and yet, surprisingly, there are few available solutions for the generation of contour plots to be used as map overlays from live data sources. The generation of outlook products on-the-fly based on live data has been made possible by advances in tools and standards for data persistence, and by a proliferation of clients with which to access standards compliant datasets. The Python based contour plot service we have developed that meets this need. The service makes use of a range of existing Python modules contributed by members of the meteorological and oceanographic communities. Using the netCDF storage format and the OPeNDAP data access protocol, our Python based web service uses Pydap to ingest data and Basemap to render the data as a contour image. The service is deployed in an Apache based environment as a Python WSGI service, and uses a HTTP interface that conforms loosely to the Open Geospatial Consortium's (OGC) Web Map Service (WMS) specification. Due to operation requirements the service contains a number of WMS extensions that limit its interoperability. Recently we have initiated a student project with the School of Applied Science at RMIT University to develop this service further, with the vision of extending the prototype already developed into a more robust, standards-compliant service that is interoperable with the range of data sources used at the Bureau of Meteorology and by the international meteorological and oceanographic communities. The integration of made-to-order contour plots from arbitrary data sources into a web based dynamical mapping tools provides opportunities for data mash-up in which data from different sources is displayed in composite. This capability to generate contour maps dynamically will open up new possibilities for interactive web-based applications for delivering forecasts and other products to users of the Bureau's services.