kisheat.py generates Google Earth overlay images and kml files
that depict your geo-located kismet output as a temperature-like
map when loaded into Google Earth.

INSTALLATION

In addition to a recent-ish python, you'll need to have the python imaging
library  (PIL) installed.  On Ubuntu flavored distros that's

sudo apt-get install python-imaging


Install the heatmap library next:

cd heatmap-1.0-kisheatpatch
sudo python setup.py install


Then copy kisheat.py to somewhere on your path of you so desire:

cp kisheat.py /usr/local/bin



USAGE

kisheat.py <[path to]Kismet-output-fileset> [essid pattern]

Only the first parameter is required.  It is not a full path of
any given file, but rather enough information to let kisheat 
know which dataset it will be working with.

Example:

Assuming that you have

Kismet-20111207-14-41-18-1.alert
Kismet-20111207-14-41-18-1.gpsxml
Kismet-20111207-14-41-18-1.nettxt
Kismet-20111207-14-41-18-1.netxml
Kismet-20111207-14-41-18-1.pcapdump


in /var/log/kismet/, you would do something like this:

mkdir output_folder
cd output_folder
kisheat.py /var/log/kismet/Kismet-20111207-14-41-18-1

Note: no file extension!  kisheat reads the  .netxml and
.gpsxml files by adding the appropriate suffix itself.


You can optionally give kisheat a regular expression to use
in deciding which networks to render.  If you do not 
provide one, all networks will be rendered.   The pattern 
can be any valid python re pattern, and will be used in the
"search" function.


kisheat has only a few options:

  -h, --help     show this help message and exit
  -l, --list     Just list xSSIDs and exit
  -a, --alldata  Include all samples taken while not moving (can exagerate
                 signal strength)


--list does what it says - it doesn't generate anything - just dumps SSIDs to stdout

--alldata requires some explanation:

The heat map rendering logic will treat repetitive mentions of the same location as an
indication that the value for that location is higher.  In practical terms this means
that if you happened to have stopped at a stop sign for a bit and you include all data
gathered there, the position at that stop sign will look like it's getting quite a lot
more signal that it really is.  So the default logic ensures that no more than
one sample is used for any given lat/lon position by selecting the first sample for 
a given point and discarding the rest.  This does *not* totally eliminate the
exaggeration effect as gps values have jitter, but it does substantially reduce it.  If you 
use the --alldata switch you disable this logic, making it likely that you will see
this effect quite prominently.



DETAILS PROBABLY WORTH KNOWING


This thing is ***SLOW***!!  If you have a really large sample, kisheat will keep your
computer busy for quite a long time.  Just let it run - the lack of verbosity 
helps speed things up.

The version of heatmap.py that kisheat uses is not pristine:  The original author 
provided a patch to it in a forum posting that allows heatmap.py to consider
weight values tied to geo-points as an indication of the "strength" of the point
as opposed to soully assuming that the number of points in a given location is 
reflective of the "strength" of that location.  The version of heatmap.py 
included here has that patch applied as well as a few other tweaks.

Note that the patched version's logic remains sensitive to repetition of
geo points - see the --alldata switch documentation for details.