This article and its child articles discuss and illustrate a possible workflow with SharpSpb.
Our goal for this exercise is to run the model on some real field data based input, then compare the model's predictions from this input set to the actual observed conditions in the field for a given number of days elapsed. We only want a crude enough output to confirm the model is behaving correctly--a true statistical analysis of accuracy would involve many more Spots and Vists. We just want to eyeball things.
Here's how we can achieve this. The Spot-Growth Database Administration Application can read a great number of different file formats representing actual Visits to actual Spots. A random selection from a big pool of data files is loaded.
That data is saved as SpotGrowth Xml (Xml data adhering to the rules defined in the SpotGrowth.xsd schema).
The data is also saved as Excel Workbook Xml, for browsing as a spreadsheet.
The SharpHog Southern Pine Beetle Infestation Model Administration Application can load SpotGrowth or SPBModel xml.
It loads the SpotGrowth xml saved previously by the SGDB Admin App.
The former field data is now formally structured input for the SharpHog model. The same data is now saved by the SharpHog SPBIM Admin App in the SPBModel.xsd format. This new file contains input for the model, but no output data has been calculated or added yet: the file is labeled to indicate that.
The SharpHog SPBIM Admin App runs the input data through the model
The SharpHog SPBIM Admin App saves the Output dataset as a differently named SharpHog xml file. One sharphog xml file contains only input, in case we want to play with it. The new xml file contains both the original input, and the resulting output as well.
We could read the SharpHog Model Output Xml in a text editor, but an easier method is to apply an Xsl transform to the Output SPBModel Xml that extracts and analyzes the numbers we are currently interested in. A transform is created and applied which outputs the results as a table in Html, ready for display in a web-page.