I have found QGIS to be an extremely powerful and extensible GIS software, in which it is possible to perform some fairly advanced data processing. One of my areas of interest is in the dynamic visualisation of model outputs, such as the progress of a flood wave across a land surface. Previously I would have used Matlab almost exclusively for this, but have recently found that in QGIS it is quicker and easier to integrate other spatial data with the visualisation. With the help of a bit of Python scripting, QGIS can animate through a series of flood output images, then spit out graphics files which can then be stitched into a movie file using software like MEncoder.
This animation was produced in such a way:
To produce this, individual graphics outputs from (a modified version of) the CAESAR-Lisflood code were loaded in tern into QGIS, overlaid on OpenStreetMap data, shaded to illustrate topography. Each visualisation image was then exported, ready to be stitched together into a video file. The visualisation code below is currently a rough work-in-progress – it would need modifying for different applications. But, I hope that it is a useful starting point for others.
Note that I modified the CAESAR-Lisflood model code to add water-source tracing capability – in this case, blue colours represent flows from the River Eden, reds are from the River Petteril and greens are from the River Caldew (see a map here). Combinations of different colours in the RGB space indicates different levels of mixing – e.g. pink/purple shows the mixing of water from the Petteril with water from the Eden. If you download the software and the accompanying Carlisle case study, you can reproduce something similar (minus the water source tracing, at present).
Prerequisites: a time-series of spatial outputs, such as those produced by CAESAR-Lisflood.
[more to follow]