createkml.py

A python script that creates a kml representation of earthquake data.

createkml.py expects a path to a csv file of the form created by the Geonet csv export call.

The implementation I use to do this reads the file downloaded by a bash script (running as a cron job)

The essential part of this is

wget -O /PATH/TO/FILE.csv "http://wfs.geonet.org.nz/geonet/ows?service=WFS&version=1.0.0&request=GetFeature&typeName=geonet:quake_search_v1&maxFeatures=50&outputFormat=csv"

The format of the file has the following columns, we are only interested in a few of them (in bold)

FID,
publicid,
eventtype,
origintime,
modificationtime,
latitude,
longitude,
depth,
magnitude,
evaluationmethod,
evaluationstatus,
evaluationmode,
earthmodel,
depthtype,
originerror,
usedphasecount,
usedstationcount,
minimumdistance,
azimuthalgap,
magnitudetype,
magnitudeuncertainty,
magnitudestationcount,
origin_geom

See http://info.geonet.org.nz/

The code

Basic stuff, no data integrity checking (this is part of a pipeline in the local implementation).

#!/usr/bin/python

import argparse
import StringIO
import csv
import datetime
import dateutil.parser
import math

#  build the parser
parser = argparse.ArgumentParser(description='Plot earthquakes on a map')
parser.add_argument('filename', help='path/to/csv file')
args = parser.parse_args()

#  open the file provided
f = open(args.filename, 'r')

#  read it in, and create the basic data we will
#  be using
output = StringIO.StringIO(f.read())

#  import in csv file format
data = csv.reader(output)

#   drop the first line
data.next()

#  print out the top of the file
print '<!--?xml version="1.0" encoding="UTF-8"?-->\n\n'
print '\tRecent earthquakes in and around New Zealand'
print '\tSource: GeoNet (www.geonet.org.nz); Created: {}'.format(repr(datetime.datetime.now().isoformat()))

# get the time right noe (UTC)
now = datetime.datetime.utcnow()

# maxdepth is the deepest value
# after which all colour values are clamped

maxdepth = 400.0

for row in data:
  # convert the date to something humans understand
  whendatetime = dateutil.parser.parse(row[3])

  # figure out how far to fade the icon
  # clamp this to ~25% minimum
  difference = (now - whendatetime).total_seconds();
  if difference &gt; 86400:
    difference = 86400
  elif difference &lt; 0:
	difference = 0

  # 'shade' is a two character hex value, 3f-ff,
  # used as the alpha value in the colour applied
  # to the icon
  shade = format(255 - int((float(difference)/86400)* 192),'02x')

  # Generate a colour value for the icon
  # associated with depth.
  # The color ramp will be a logarithmic
  # progression from 00ff00 (shallow)
  # to 0000ff (deep), maxed at maxdepth km?
  depth = float(row[7])

  # fast at shallow, slow at depth
  # multiplier = 255/sqrt(maxdepth)
  # depth2 = math.sqrt(depth)*(255/math.sqrt(maxdepth))

  # slow at shallow, fast at depth
  # divisor = maxdepth/sqrt(255)
  # depth2 = (depth/(maxdepth/math.sqrt(255)))**2

  # flat
  # multiplier = maxdepth/255
  depth2 = depth*(255/maxdepth)

  if depth2 &gt; 255:
    depth2 = 255.0
  elif depth2 &lt; 0:
    depth2 = 0.0
  if depth2 &lt; 128:
    green = 'ff'
    red = format(int(depth2*2),'02x')
  else:
    green = format(255-int((depth2-128)*2),'02x')
    red = 'ff'

  # round magnitude to determine the icon style
  magnitude = float(row[8]);
  magnitudelow = int(math.floor(magnitude))
  if magnitudelow &gt; 9:
    magnitudelow = 9
  elif magnitudelow &lt; 0:
    magnitudelow = 0

  description = '<![CDATA[\n
<div style="text-align: left;">'
  description += '<a href="http://www.geonet.org.nz/quakes/region/newzealand/' + row[1] + '" target="' + row[1] + '"><b>' + row[1] + '</b></a>'
  description += '\n
<b>' + whendatetime.strftime('%H:%M, %d %B %Y') + '</b> (UTC)'
  description += '\n
magnitude: <b>' + str(round(float(row[8]),2)) + '</b>'
  description += '\n
depth: <b>' + str(round(float(row[7]),2)) + 'km</b>'
  description += '</div>
]]>'

  print '\t'
  print '\t\t' + description + ''
  print '\t\t' print '\t\t\n\t\t\t' + row[6] + ',' + row[5] + ',0' print '\t\t\n\t' # print out the bottom of the file print '\n\n'

The script dumps its output to standard output, so I generally redirect this to a file.