Why

lets back up a minute ...

Why Program?

What is wrong with ...

This?

well nothing actually ...

... until you hit a deadline

... and realize you made a mistake

... in one of the intermediate analysis steps

... three weeks ago

Because it is not a GIS tool

It is used in

• Web frameworks
• Data Visualization
• Image Processing
• Numerical Modeling
• Signal Processing
• Machine Learning
• Big Data
• Blah Blah Blah

Oh and also in GIS

PyProj (PROJ.4)

• generalized projection engine
• projections determined by: proj4 string or EPSG number
• UTM14

  EPSG number: 32614
  proj4 string: '+proj=utm +zone=14 +ellps=WGS84 +datum=WGS84 +units=m +no_defs'

• http://spatialreference.org, http://www.epsg-registry.org/

Read csv data from url and project Travis County Cities to UTM14

"""
url - http://goo.gl/WFylXY
data format: 
  "zip_code","latitude","longitude","city","state","county"
  "00501",40.922326,-72.637078,"Holtsville","NY","Suffolk"
  "00544",40.922326,-72.637078,"Holtsville","NY","Suffolk"
  ...
"""
import pyproj
import pandas as pd
url = "http://goo.gl/WFylXY"
data = pd.read_csv(url)
travis_cities = data[data['county']=='Travis']
lons = travis_cities['longitude'].values
lats = travis_cities['latitude'].values
utm14 = pyproj.Proj("+init=EPSG:32614")
travis_cities['utm_x'], travis_cities['utm_y'] = utm14(lons, lats)
          

GDAL and OGR

fiona and rasterio

Standards are like toothbrushes, everybody agrees you should have one, but no one wants to use yours. — Joe Croser

import pandas as pd
from shapely.geometry import Point, mapping
from fiona import collection
 
schema = { 'geometry': 'Point', 'properties': { 'city': 'str', 'zip': 'str' } }
url = "http://goo.gl/WFylXY"
data = pd.read_csv(url)
with collection("zipcodes.shp", "w", "ESRI Shapefile", schema) as output:
    for index, row in data.iterrows():
        point = Point(row['longitude'], row['latitude'])
        output.write({
            'properties': {'city': row['city'], 'zip': row['zip_code']},
            'geometry': mapping(point)
        })
          

ftp://rockyftp.cr.usgs.gov/vdelivery/Datasets/Staged/NED/1/

print 'Downloading tiles needed for requested bounding box:'
raster_tiles = []
tiles = _tile_urls(layer, xmin, ymin, xmax, ymax)
for i, url in enumerate(tiles):
    filename = os.path.split(url)[-1]
    zip_path = os.path.join(path, layer_dict[layer], 'zip', filename)

    print '... downloading tile %s of %s from %s' % (i+1, len(tiles), url)
    util.download_if_new(url, zip_path, check_modified=True)
    print '... ... zipfile saved at %s' % zip_path
    tile_path = zip_path.replace('/zip', '')
    raster_tiles.append(_extract_raster_from_zip(zip_path, tile_path))

print 'Mosaic and clip to bounding box extents'
tile_path = os.path.split(tile_path)[0]
print subprocess.check_output(['gdalbuildvrt', '-te', repr(xmin), repr(ymin), repr(xmax), repr(ymax), output_path] + raster_tiles)
return output_path
        

xmin, xmax, ymin, ymax = (-79.68433821243661, -77.42742156945509, 
        32.81417227179455, 34.998064728936754)
filename = get_raster('1 arc-second', xmin, ymin, xmax, ymax)
with rasterio.drivers():
    with rasterio.open(fname) as src:
        dem = src.read_band(1)
        ndv = src.nodatavals
dem[dem==ndv] = np.nan
plt.imshow(dem)
plt.savefig('dem.png')
        

AEIDW - Anisotropic Eliptical Inverse Distance Weighting