Weather Data Visualization

Project 3, CS 526
April, 2004

Ratko Jagodic

Written entirely in C++ using FLTK/FLUID for GUI. No threads, all the GUI events and display
updates are handled manually for each control.

Source files and binaries: weather.zip (includes source code, FLTK/FLUID files and a WIN32 executable)

How to run:
The executable can be run on any Windows machine that has VTK installed and appropriate paths
setup (environment variables) for the libraries and dlls.
Also, when running the executable make sure that the data file and all the images are in the same
directory as the executable.

Features:

• Different types of data:
  • Average temperature (colormaps)
  • Average soil temperature (colormaps)
  • Average precipitation (colormaps)
  • Average humidity (colormaps)
  • Average wind speed and direction (colormaps and vectors
• Every type of data has it's own lookup table so that the different visualizations
  are easily recognizable. They also try to follow some standard convention that we
  are all used to (i.e. red is hot, blue is cold). In every visualization darker means more
  of that data (i.e. more humidity, precipitation, wind...)
• The included dataset is for March 2003 but any dataset could be loaded since nothing
  has been added or deleted from the original file provided by WARM
• All of the data has been interpolated using the built in vtkShepardMethod.
• Different interpolation smoothness can be adjusted via GUI.
• Also, the memory requirement is very limited because just the data for the required day is
  loaded at any point. This actually doesn't affect the performance that much either.
• The wind direction is visualized using vectors (arrows) and it can be layed on top of any other
  visualization.
• The wind speed has been visualized using colormaps. That seemed to be more informative
  since it is easier to distinguish different colors than different lengths of a whole bunch of vectors.
• Whenever visualization style has been changed (or the day itself), a range of values is displayed
  in the GUI.
• In order to put all this data into some context, 5 different base maps of Illinois have been provided.
  • Roads
  • Interstates
  • Counties
  • Water supply (when looking at precipitation data, drought and such)
  • Major towns

Features not behaving correctly:

• No streamlines. Honestly, I had no idea that we even had to visualize wind using streamlines
  as well. I thought we had a choice of vectors or streamlines or whatever method we think is
  the best. Well, I thought vetors are sort of a natural mapping for visualizing wind data so I haven't
  even attempted to visuzalize it using streamlines.
• I would like to have something like vtkScalarBar that provides a legend of the colors on the screen
  instead of just max and min values.
• Picking. I have tried that however, after spending some time on it I was not able to extract the correct
  scalars. It extracts the point ids fine but when I do GetValue on the PointData, it gives me odd values.
  (see image below).
  

<< Back to main page

Screenshots:

Figure 1. Just a regular average air temperature visualization. No smooth interpolation. That seems to
show the boundries of temperature changes much better. It also looks more like the weather forecast
maps.

Figure 2. Temperature data with smooth interpolation. it looks prettier but less functional in my opinion.
Note, the underlying map has changed. and the wind direction has been added as well.

Figure 3. Same as above. Same date except there is not wind and the smoothing of the data is not on.
It shows the variations in temperature a lot better.

Figure 4. Soil temperature visualization with smoothing on. Darker means lower temperature. The
color lookup table is different than the air temperature one on purpose (to better distinguish the two,
also green sort of represents grass or ground).

Figure 5. Precipitation data with yet another base map. Also, the lookup table has been made especially
for this visualization so that it represents water better. White means no precipitation and darker means
more and more precipitation. Smoothing is off.

Figure 6. Again, different map with wind visuzalization. Wind speed has been visualized as a colormap
and wind direction as vectors. Darker means greater wind speed.

Figure 7. Humidity visualization with wind direction over it. The same day as Figure 5. Notice that
in the figure 5. it is raining in most of the state, hence, more humidity.

Figure 8. This image shows a (non working) picking example. In the lower left corner you can see
the output as different points have been picked. The point Ids have been extracted correctly but the
scalars have not.