In [1]:
%matplotlib inline
Mapping the Republic of Letters (Stanford) https://web.stanford.edu/group/toolingup/rplviz/
Franklin and Voltaire Correspondence Winterer, "Where is American in the Republic of Letters," http://dx.doi.org/10.1017/S1479244312000212
Million-dollar blocks http://spatialinformationdesignlab.org/projects.php%3Fid%3D16
Spatial imagination https://web.stanford.edu/group/spatialhistory/cgi-bin/site/pub.php?id=93
Digital Harlem http://digitalharlem.org/
For detailed tutorials for historians, see Programming historian http://programminghistorian.org/lessons/#mapping-and-gis
Cartography not value neutral.
Mathematical choices have geopolitical significance.
See the classic work of David Harvey, The New History of Maps
to get intuitions:
how many ways to project a sphere onto a flat surface? Is Greenland bigger than Africa in your choice of projection? Is that ok?
key python package geopy
not by default in conda; you can get with pip install geopy.
We'll just use requests as the geocoding APIs are straightforward.
In [2]:
import requests
payload = {'address': 'Columbia University'}
r = requests.get("https://maps.googleapis.com/maps/api/geocode/json", params=payload)
#cf.:
#https://maps.googleapis.com/maps/api/geocode/json?sensor=false&address=Columbia+university
#NOTE you'll need an authentication key
#so that Google doesn't get testy with you and stop giving results
In [3]:
r.json()
Out[3]:
In [4]:
r.json()["results"][0]['geometry']['location']
Out[4]:
FYI:
geopy just makes this a little easier:
Once you've intalled geopy, to initialize the geocoder:
from geopy.geocoders import GoogleV3
my_google_key="<PUT YOURS HERE>"
geolocator=GoogleV3(api_key=my_google_key)
Then to query it's simply:
location=geolocator.geocode("<address here>")
or for address, lat and lon,
address, lat, lon=geolocator.geocode("<address here>")There are alternatives to the Google. OpenStreetMap's nominatim is an important example.
In [5]:
payload = {'q': 'Columbia University', 'format':'json'}
r = requests.get("http://nominatim.openstreetmap.org/search/q", params=payload)
In [6]:
r.json()
Out[6]:
In [7]:
r.json()[0]['lon']
Out[7]:
In [8]:
lon=r.json()[0]['lon']
lat=r.json()[0]['lat']
actually they need to be floats!
In [9]:
lon=float(r.json()[0]['lon'])
lat=float(r.json()[0]['lat'])
In [10]:
##Let's start with a list of places
places=["Reno, NV", "Somerville, MA", "New York, NY"]
In [11]:
lons=[]
lats=[]
for place in places:
payload = {'q': place, 'format':'json'}
r = requests.get("http://nominatim.openstreetmap.org/search/q", params=payload)
lon=float(r.json()[0]['lon'])
lat=float(r.json()[0]['lat'])
lons.append(lon)
lats.append(lat)
##FOR REAL CODE BUILD IN TRY....EXCEPT !!
In [12]:
lons
Out[12]:
Lots of packages, low and high level.
http://carsonfarmer.com/2013/07/essential-python-geo-libraries/
vincent and bokeh.Check them out; they are beautiful and easy to use.
We'll use the extension to matplotlib for contending with geographical data called basemap.
Basemap tutorials: https://basemaptutorial.readthedocs.org/en/latest/index.html
Builds on the full rich matplotlib world.
In [13]:
#first our imports
#may need to conda install basemap
from matplotlib import pyplot as plt
from mpl_toolkits.basemap import Basemap
In [14]:
# set up orthographic map projection with
# perspective of satellite looking down at 50N, 100W.
# use low resolution coastlines.
map = Basemap(projection='ortho',lat_0=45,lon_0=-100,resolution='l')
# draw coastlines, country boundaries, fill continents.
map.drawcoastlines(linewidth=0.25)
map.drawcountries(linewidth=0.25)
map.fillcontinents(color='coral',lake_color='aqua')
# draw the edge of the map projection region (the projection limb)
map.drawmapboundary(fill_color='aqua')
# draw lat/lon grid lines every 30 degrees.
map.drawmeridians(np.arange(0,360,30))
map.drawparallels(np.arange(-90,90,30))
plt.title('USA')
plt.show()
In [15]:
#now use with "blue marble visualization"
map = Basemap(projection='ortho',lat_0=45,lon_0=-100,resolution='l') #not asking for high resolution!
# draw coastlines, country boundaries, fill continents.
map.drawcoastlines(linewidth=0.25)
map.drawcountries(linewidth=0.25)
# draw the edge of the map projection region (the projection limb)
# draw lat/lon grid lines every 30 degrees.
map.drawmeridians(np.arange(0,360,30))
map.drawparallels(np.arange(-90,90,30))
map.bluemarble()
plt.title('USA')
plt.show()
In [16]:
#Basic Map
fig = plt.figure(figsize=(15,12))
#choose projection
projection='merc' #Mercator projection
#choose the rectangle we're interested in seeing by picking the left lower and right upper corner
x1 = -135.
x2 = -60.
y1 = 25.
y2 = 55.
# now initialize the map
m = Basemap(projection=projection, llcrnrlat=y1, urcrnrlat=y2, llcrnrlon=x1,
urcrnrlon=x2, resolution="i")
# choose features to visualize
m.drawcoastlines()
m.drawmapboundary()
m.drawcountries()
m.fillcontinents(color = '#CCCCCC')
#finally plot
plt.show()
In [17]:
#start the same
fig = plt.figure(figsize=(15,12))
projection='merc' #Mercator projection
x1 = -135.
x2 = -60.
y1 = 25.
y2 = 55.
m = Basemap(projection=projection, llcrnrlat=y1, urcrnrlat=y2, llcrnrlon=x1,
urcrnrlon=x2, resolution="i")
m.drawcoastlines()
m.drawmapboundary()
m.drawcountries()
m.fillcontinents(color = '#CCCCCC')
#now add the points to plot! the object m will convert lon, lat into coordinates to plot!
x,y=m(-71.0994968, 42.3)
m.plot(x,y,marker='D', color='m')
plt.show()
In [18]:
# now let's pass our list
fig = plt.figure(figsize=(15,12))
projection='merc' #Mercator projection
x1 = -135.
x2 = -60.
y1 = 25.
y2 = 55.
m = Basemap(projection=projection, llcrnrlat=y1, urcrnrlat=y2, llcrnrlon=x1,
urcrnrlon=x2, resolution="i")
m.drawcoastlines()
m.drawmapboundary()
m.drawcountries()
m.fillcontinents(color = '#CCCCCC')
#now add the points to plot! the object m will convert a list of lon, and of list of lat into coordinates to plot!
x,y=m(lons, lats)
m.plot(x,y,marker='D', color='m')
plt.show()
Alternatively, for a simple example of combining basemap with networkx, see the answer code in https://stackoverflow.com/questions/19915266/drawing-a-graph-with-networkx-on-a-basemap
use scatter instead to plot lists of lat, lon
In [19]:
# once again, with feeling. or rather, with scatter
fig = plt.figure(figsize=(15,12))
projection='merc' #Mercator projection
x1 = -135.
x2 = -60.
y1 = 25.
y2 = 55.
m = Basemap(projection=projection, llcrnrlat=y1, urcrnrlat=y2, llcrnrlon=x1,
urcrnrlon=x2, resolution="i")
m.drawcoastlines()
m.drawmapboundary()
m.drawcountries()
#now add the points to plot! the object m will convert lon, lat into coordinates to plot!
x,y=m(lons, lats)
m.scatter(x,y,150,marker='o', color='m')
plt.show()
See https://basemaptutorial.readthedocs.org/en/latest/index.html for how to do each of these things
and for inspiration
see Nathan Yau, Visualize This, ch. 8 "Visualizing Spatial Differences", on-line for CU folks at http://www.columbia.edu/cgi-bin/cul/resolve?clio9683912
In [22]:
fig = plt.figure(figsize=(15,12))
projection='merc' #Mercator projection
x1 = -135.
x2 = -60.
y1 = 25.
y2 = 55.
m = Basemap(projection=projection, llcrnrlat=y1, urcrnrlat=y2, llcrnrlon=x1,
urcrnrlon=x2, resolution="i")
m.drawcoastlines()
m.drawmapboundary()
m.fillcontinents(color='#ddaa66',lake_color='aqua')
m.drawcountries()
##add the states
m.drawstates()
plt.show()
In [21]:
from IPython.display import Image
Image("https://www.msu.edu/~dougl126/images/cancer_rates_jenks.jpg")
Out[21]:
We make data from sources: we don't find it pregiven.
Data is made, not born: fully artificial
artificial therefore false
Or artificial therefore way to create something positive
Not positivist social sciences, not social sciences controlled by computation but social sciences with computation tools, warts and afforances alike
In [ ]: